Exploring New Project Ideas

RE: TMF Ontology SIG [#169800]

pierregauthier — Wed, 14 Sep 2011 18:03:16 GMT

Yes I will create the community projet this week.

Pierre

From: Exploring New Project Ideas [mailto:newprojects@community.tmforum.org] On Behalf Of Stephen Estes
Sent: Wednesday, September 14, 2011 1:25 PM
To: pierre.gauthier@osswave.com
Subject: [Exploring New Project Ideas] TMF Ontology SIG [#169800]

Stephen Estes , Business Process Architect , Cisco Systems

Hello, Pierre -

Is the new Ontology SIG community moving forward?

Thanks,

-- Stephen

[collapse]On Aug 23, 2011, at 8:22 PM, Pierre Gauthier wrote:

Exploring New Project Ideas
Topics:General

Shared In:4 Communities

Shared In:Information Framework (SID) : General
Business Process Framework (eTOM) : General
TM Forum Integration Program (TIP) : General
Frameworx Architecture Steering Team (FAST) : General

View thread online

Please note, this discussion thread is shared across communities. Any replies that you make to this discussion thread, will be posted to all of the communities and topics listed above.

r>

Full thread available online here:
http://www.tmforum.org/community/groups/information_framework_sid/forum/p/169800/192332.aspx?utm_campaign=community&utm_medium=information_framework
_sid&utm_source=alert#192332

----------------------------------------------------------------------------------------------------------------------------------------------------
This message was sent to dingersoll@versant.com and follows the subscription preferences of your membership of the Information Framework (SID) TM
Forum Online Community.

To unsubscribe immediately, please click here and click the 'Edit Subscriptions' button from the Email Subscription preferences section located on
the right hand side of the page just underneath the menu bar. To cut down on email volume, you may also choose to receive your discussion
notifications via a daily or weekly digest email as opposed to immediate separate emails.

TM Forum, 240 Headquarters Plaza, East Tower, 10th Floor, Morristown, NJ 07960-6628, USA. Call +1 973 944 5100. info@tmforum.org

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.

8/16/2011 5:54:44 PM

"However, at the current time, it would be premature to suggest that the actual models (and a couple of the transformation engines) would be contributed to the TMF as these component parts currently fall into one of two categories; either proprietary work of a consortium, or proprietary work of a individual/corporation."

Ah!

Would be most interested in TMFs opinion on this as a Forum/Consortium on this point. I agree with Joel, we say no this is not going to happen for various reaons and do other less ambitious but yet useful things as a group, or we say yes we can make this happen and put in the time and effort to come up with work. (New IPR though i understand very little of that document ) it sounds is intended to encourage the latter sort of thing?

Sitaraman

Stephen Estes , Business Process Architect , Cisco Systems

8/16/2011 5:48:11 PM

OWL 2.0 is a current standard, and its subset, OWL-DL is both very expressive and logically tractable. For this, the editing tool, Protege 4.1x, is open source, multi-platform (Java) and robustly maintained.

http://protege.stanford.edu/

-- Stephen

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.

8/16/2011 5:46:19 PM

And yes, as Ken says one group can feed into the other so the whole becomes greater than the parts with team focused on clearly defined agendas and work in the same environment so that there not avodable software-administrrative hassles that hurt progress. In that sense a Metamodel is indeed an important first step, we need to have our grammar straight before we use it to consistently communicate semantics without chaos due to misinterpretation.

Sitaraman

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.

8/16/2011 5:42:11 PM

Dave/Jerome,

Yes, I would go with any set of deliverable as long as it is useful to the industry and those are clearly defined.

Having said that, i do feel that in a forum called Ontology SIG, the participants who expressed interest in this group as long as such a group exists should be allowed to go beyond a meta model.

A meta model, to me is the equivalent of an XSD. I helps validation, it checks for the grammar for the language spoken is correct, the statement could very well be to borrow from a poem,

It would be perfectly valid (context awareness, i am reminded of John Strassners post) to say "would like a footlong with everything on it please" in a blockbuster store. So we are back to the killed thread i am afraid, context aware policies and behaviors. But even that belongs in a thread for PBNM or PDNM,

But for a thread that calls itself Ontology SIG should be allowed to explore beyond policies and contexts. even.

Therefore a

a)SID metamodel/Interface Metamodel group can have as a deliverable to come up wth XSD equivalent of Models and Interfaces.for which BTs metamodel doc could be a base. Additional deliverables, automated consistency checks, both static when a new model entities or interfaces are defined(similar to strong typing carrier over to a higher layer here) up to fine grained interface signature/parameter level dynamic checking. We could even extend this to automatic translation between instances of two diferent meta models. About here the name justifies a stop.

b) PBNM which will deal with how to express policies and behaviours an context awareness and determining what are the typical policies that are useful.

c) Onotology SIG, where we go beyond metamodels and context awareness is there a fundametnal structure to knowledge representation, are there fundamental operations on those which would formalize knowledge evolution (collective knowledge, learning). I like Pauls definition, though it could turn out to be ambitious with a narrower focus to our domain by trading off loss in generality, we would have a surer sight of the goal.

Additional deliverables would be,

Exhaustively list concrete Network Semantic Instances.

Map problem space in telecommunication network operations to NEtwork Semantics

Specify using knowledge evolution as above for

Interface harmonization without loss of semantics

Simple RCA/SIA

Self-*

Trending and Planning.

If the feeling is above is too much to shoot for, i will go with it, but the group should cease to be called Ontology SIG, Ontology and its potential applications to Telecommunications Network
go beyond metamodeling is my feeling.

Sitaraman

Joel Fleck II , Chief Architect , Hewlett-Packard

8/16/2011 5:41:58 PM

Wedge et al:

One of the key directions in research and industry looking forward is the use of a model-driven approach for the construction of business vocabularies, rules, ontologies and technology-neutral representations of the models. In particular, recent research has demonstrated a very usable approach using the OMG’s Semantics for Business Vocabulary and Rules (SBVR) as the basis for transformations to OWL 2.0 (for ontologies and ontological based reasoning –read that semantics), and as the basis for N3 or N-tuple technology-neutral based representations for the policy language(s) of choice, to UML and OCL for the generation of code stubs, and finally to XPATH 2.0 for testing potential interfaces. For reference, I refer to a paper in IEEE Communications August, 2010 “Multidomain IT Architectures for Next-Generation Communications Service Providers”.

Just as an FYI, there has been significant effort on the formulation of an SBVR core for the specification of inter/intra domain relationships and the management of these relationships (and the transformations of that work into UML, OWL, and N3). The methodology has been successfully proven and shown to be very valuable in capturing the business perspective and building software architecture models from them. I would strongly advise investigation into these approaches.

However, at the current time, it would be premature to suggest that the actual models (and a couple of the transformation engines) would be contributed to the TMF as these component parts currently fall into one of two categories; either proprietary work of a consortium, or proprietary work of a individual/corporation.

Best Regards,

Joel

Joel J. Fleck, II
Chief Architect

Head of Standards

HP Software CTO
Hewlett-Packard Company
200 Connell Parkway, Suite 500

Berkeley Heights, NJ 07922

+1 908.898.4091 Phone
+1 908.416.1706 Mobile
+1 908.898.4887 Fax
joel.fleck@hp.com

Kenneth Dilbeck , Director, Technical Program , TM Forum

8/16/2011 5:16:33 PM

Folks,

Anyone familiar with SADL http://sadl.sourceforge.net/?

Perhaps if both groups can use the same tool base, Eclipse, the chances of keeping them together would increase. The JOSIF work is Eclipse based as well.

It would seem having tool support for something as rigorous as an ontology would be helpful. I have only taken a quick look and this may be more oriented towards reasoning systems but thought I would ask.

We do need to make sure there is a very open line of communication between the FAST meta-model team and any other ontology work that takes places to avoid duplication effort and development of conflicting answers.

Regards,

Ken

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.

8/16/2011 4:47:53 PM

Hi Wedge,

Isnt modelling itself an abstraction activity, therefore 2 and 1 are same?

Sitaraman

William Greene , Technical SME; CTR DES-E , DOD

8/16/2011 4:46:08 PM

Classification: UNCLASSIFIED
Caveats: NONE

I’m a little shy of using “automated reasoning” as a goal. I’m personally satisfied with “automated services” as a more reachable goal. Perhaps, "automated, adaptive services” as a stretch goal.

I'm simply happy to start with defining the grammar, syntax, semantics. Also, maybe a few basic patterns that support first understanding Frameworx as a model and then basic operations on information contained in the model.

Eventually, (but not soon ;-) we might get to a point where we can use this to, for example, “frame the set of available interface solutions for domain X and iterate across these to determine an optimal decision tree for function L: federate with the similar decision tree coming from Domain Y.

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

David Milham , Chief Architect , TM Forum

8/16/2011 4:30:55 PM

One needed to be on the call to hear the full background and at the presentation at TAW Baltimore

The key point made by Steve was that this was an example of an enterprise repository solution some aspects of which need to be embedded in Frameworx, Without such a metamodel SP like BT assert that is it difficult to practical evolve systems interfaces business services etc. And it is not primarily a technical issue it is a management and governance need on Frameworx.

Dave

Dave Milham

Chief Architect

t +44 5601 855561 m+44 7802 902451
edmilham@tmforum.org

Paul Hanke , Sr Professional Staff , Johns Hopkins University Applied Physics Lab

8/16/2011 4:15:48 PM

... I'm not saying that this is what everyone thinks of when they hear the word "ontology", but here is what a deliverable could look like:

An extensible, machine-readable, and foundational logical theory of service provider operations that supports automated reasoning and automated interoperability even across arbitrary domain-specific extensions.

That is, unlike a data model expressed in XML Schema where the best a machine can do with instance data is simply ingore any extensions it is unfamiliar with, an ontology expressed in OWL opens the possibility that a machine can formally manipulate instance data (in a way that mimics human understanding) to include extensions that it has never seen before ... perhaps this is what the FAST team is trying to address with meta-models, I couldn't say ... but even if it is, at this formative stage I see room for both thrust areas to poke and prod at the envelope of the possible.

William Greene , Technical SME; CTR DES-E , DOD

8/16/2011 4:15:42 PM

Classification: UNCLASSIFIED
Caveats: NONE

From Stephen, " Also, the JOSIF open source project has define and implemented a meta-model for interfaces."

Brilliant work!

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

David Milham , Chief Architect , TM Forum

8/16/2011 4:08:59 PM

Steve

Correct thanks for reminding us of the location of that information

Dave

Dave Milham

Chief Architect

t +44 5601 855561 m+44 7802 902451
edmilham@tmforum.org

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.

8/16/2011 4:06:43 PM

"David Milham" wrote:

Hi Dave,

Need a favor,

The following document seems to be a MetaModel proposal from Steve Orobeck, this looks like more of a solution document than a list of problems bothering a service provider as you seem to have intended. Am i looking at the right document and if so a pointer to where the problem statements are would help.

If you want to see what is bothering at least one service provider you should look up this contribution.

http://www.tmforum.org/Community/groups/frameworxarchitecture/contributions.aspx?id=artf2509

which was originally presented at TAW Baltimore in the cross Frameworx meetings

The following document gives me a permission denied error, would much aprpeicate if it can be sent as an attachment.

Thanks much.

Sitaraman

http://collab.tmforum.org/integration/viewcvs/viewcvs.cgi/trunk/?root=business_services&system=exsy1002

but we discovered that the individual meta-models don’t line up as we would wish.

Hence we started a top down exercise to complement the prior approach.

What we need now is hand and bodies to work through the patterns we need to use and identify and rectify the gaps in the existing component frameworks.

Dave

Dave Milham

Chief Architect

t +44 5601 855561m+44 7802 902451
edmilham@tmforum.org

From:Information Framework (SID) [mailto:sid@community.tmforum.org] On Behalf Of Sitaraman Wills
Sent: 16 August 2011 19:05
To: Dave Milham
Subject: [Information Framework (SID)] TMF Ontology SIG [#169800]

Sitaraman Wills, Senior Software Engineer , NMSWorks Software Pvt. Ltd.

At the end fo the day, we are TMF first and then Ontologists (to various levels of expertise, from bystanders, through believers, via budding through to been-there-done-thats ).

Perhaps we must have another line of attack from the ServiceProviders side. We could ask operators running their networks day in and day out...tell me in plain english or you native language what are the things that bother you in your job.

There must be complete freedom in responding, no tracking along a direction with a set of 10 questions and the like.

Then we huddle together and extract semantics out of them. Then, we can have that comfortable feeling that, at the end of the day, what we come up with is going to make an important set of people that drive an all pervading industry, smile, courtesy TMF.

Sitaraman

Information Framework (SID)
Topics:General

Shared In:4 Communities

William Greene , Technical SME; CTR DES-E , DOD

8/16/2011 4:02:35 PM

Classification: UNCLASSIFIED
Caveats: NONE

I think we here and also the FAST team needs to distinguish between meta-model as:
- a model of a model
- an abstraction of a model.

At least in ontology, it should be an abstraction of a model.

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

Stephen Fratini , Senior Engineer , Telcordia Technologies

8/16/2011 3:40:13 PM

Dave and All,

Also, the JOSIF open source project has define and implemented a meta-model for interfaces.

See the JOSIF wiki pages at

http://sourceforge.net/apps/mediawiki/openoss/index.php?title=Interface_Meta-Model

Steve

- Stephen

William Greene , Technical SME; CTR DES-E , DOD

8/16/2011 2:24:04 PM

Classification: UNCLASSIFIED
Caveats: NONE

Yea, Jerome, not so easy ;-)

These are characterizing statements; perhaps true; but neither providing definitions nor establishing relationships.

For example, Behavior and Lifecycle, must be defined in the semantics of the Ontology. The metamodel executes a ontological syntax, a grammar, in applying behavior and Lifecycle to specific targets. (first pass)

Sitaraman, I think, raises an interesting supposition: is the Frameworx metamodel recursive or self referential? Should that be a structural requirement?

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.

8/16/2011 1:57:06 PM

At the end fo the day, we are TMF first and then Ontologists (to various levels of expertise, from bystanders, through believers, via budding through to been-there-done-thats ).

There must be complete freedom in responding, no tracking along a direction with a set of 10 questions and the like.

Sitaraman

Jerome Hannebelle , Director, Open Innovation, FT/IMG/Technocentre/BP , France Telecom

8/16/2011 1:49:04 PM

Model experts are certainly more qualified than I, but I would think :

1) While an ontology focuses on giving a meaning(= a semantic) to "objects" (sometimes not as formally as a metamodel), a metamodel probably goes beyond in also looking at behavior, lifecycle, ...

2) The metamodel may define abstract, unifying concepts that may not show up in an ontology (because they are abstract).

Mr experts, please jump in !

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.

8/16/2011 1:34:32 PM

I am particularly interested (from the most generatl interface harmonization perspective) in the following goals.

- Can we structure Semantics (we need not consider the general programming language semantics, a perhaps more difficult problem, but can confine to the most frequently occuring, to begin with, semantic patterns in telecommunication networks) Sub goals would be then

a) What are the frequently occuring "network semantics" both across and within Networks. We can and perphaps must free ourselves of all constraints and feel free to to be abstract here. Then

b) Can we represent them in COMMON structural pattern, concretely well defined with as little ambiguity as possible.

- Now that we have a COMMON pattern, for the most popularly occuring (to keep things simple at first) concretizations seen above

a)Represent different instances of the same concrete Semantics in the common pattern and lay them out.

b)In the abstract (free form) how do we characterize the "extent of difference" between the two.

c)How do we concretize the differnce expression (procedures, algorithms on the said common structure) that unambigously characterizes the diff. That is, from the diff, one must be able to get to the other from the other, other and converse.

d) If we do reach c) then hopefully the holy grail of merge (All of A in B, or All of B in A or B = A+B or B = B-A and other possibilities) will be a simpler extension.

I am hoping, well i would go the extent of saying it is almost a hunch, that the common strcture would have aspects of self-simlarity. Therefore i am hoping the above operations can be recursively applied to whatever turns out to be the "Atomic Operation" on the said structure and therefore be able to do complex "semantic merges", "Semantic Morphs" or Semantc Overlays and such operations out of the recursive operations composed out of the atomic operations.

If we reach here we have set ourselves up to casting these a)Structure and b)Unit )perations in telecommunicational semnatic processing (due credits to the Chemical Engineering clan),

Then if only we get lucky and characterize state (a poorer cousin of full semantics perhaps) with the same pattern, if state is a simpler form of semantics it does seem likely,

-Record State of network entities, of groups of network entities (same state compositional types as above, we can reuse those)

- Give said state(s) as input to Unit Operations and its constituents, and based on the input call suitable child operations.... We have context aware, semantic processing,

- Now have state change as part of the child operation and there you have feedback control, with semantics and syntax, a partial learning system, now we can appeal fo more higher levels of concretized math, Stability, Pole-Zero Analysis and other such intriguing things with which Control Theorists work their systems....?

And here it is... I would much appreciate peoples time in responding why that common structure cannot be the simple TREE of arbitray depth and in general a system represented by a forest of such trees? Concrete examples where it fails in achieving a given functionality would help much

Thanks.

Sitaraman

William Greene , Technical SME; CTR DES-E , DOD

8/16/2011 1:23:22 PM

Classification: UNCLASSIFIED
Caveats: NONE

Jerome, could you expound on " believe relates to FAST team's intent to define the Frameworx's Metamodel." My gut feeling is you are right (that is, I am not disputing this comment at all). I feel this is an important statement; but I need some help seeing exactly how the ontology and the metamodel are linked.

From: Wedge Greene

Craig Gallen , Consultant OpenNMS group , The OpenNMS Group, Inc.

8/16/2011 1:08:15 PM

Hi Pierre,

As discussed, I am very interested in this and would like to include the ten University labs I represent through the web Science trust. www.webscience.org

Cheers

Craig

Jerome Hannebelle , Director, Open Innovation, FT/IMG/Technocentre/BP , France Telecom

8/16/2011 1:06:36 PM

Hi Pierre,

This is definitely an interesting topic, which I believe relates to FAST team's intent to define the Frameworx's Metamodel.

Jerome

William Greene , Technical SME; CTR DES-E , DOD

8/16/2011 1:02:58 PM

Classification: UNCLASSIFIED
Caveats: NONE

I'm interested. However, I think after establishing a basic syntax and some inclusive semantic patterns (e.g., search, aggregation, collaboration, policy), fleshing out the various semantic definitions can be handled in the many traditional workgroups. What we eventually need is for each team to feed their semantics into the overall ontology just like we currently feed the SID and eTOM.

My specific interest today is on how an ontology will work with policy.

I do not see we need to get it perfectly right the first time; just we get it close enough to do much of the work... old 80/20 rule.

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.

8/16/2011 12:46:06 PM

Hi Pierre,

Meself in it, initially however i request i be granted aceess to discussions and confine myself to observing and learning by participitating in those discussions. I do hope that would be ok.

Thanks.

Sitaraman

Alex Zhdankin , Modeling Architect , Cisco Systems

8/16/2011 12:42:30 PM

Count me in.

Alex.

John Wilmes , Chief Technical Architect, Communications Sector , Progress Software

8/16/2011 12:27:07 PM

I would like to participate especially with reference to SID and eTOM.

John

Paul Hanke , Sr Professional Staff , Johns Hopkins University Applied Physics Lab

8/16/2011 12:27:00 PM

... responding with interest :-)

Bob Natale , Chief Engineer, Joint NetOps & Cyber Defense , MITRE

8/16/2011 12:16:44 PM

Count me in, Pierre.

Pierre Gauthier , CEO and Principal Consultant , Oss Wave

8/16/2011 11:50:18 AM

I would like to see if there is enough interest for creating an Ontology SIG within TMF.

I sugest that the initial focus be on establishing the foundation for a TMF Frameworx Ontology.

It is important that the set of proposed ontologies provides complete coverage of the Frameworx. This involves conceptual modeling of eTOM, SID, TAM but also complete coverage of the Business Services and TIP.

Please reply to that post if you are interested in participating to the TMF Ontology SIG.

End of Discussion History

Full thread available online here: http://www.tmforum.org/community/groups/exploring_new_project_ideas/forum/p/169800/193442.aspx?utm_campaign=community&utm_medium=exploring_new_project_ideas&utm_source=alert#193442

This message was sent to pierre.gauthier@osswave.com and follows the subscription preferences of your membership of the Exploring New Project Ideas TM Forum Online Community.

To unsubscribe immediately, please click here and click the 'Edit Subscriptions' button from the Email Subscription preferences section located on the right hand side of the page just underneath the menu bar. To cut down on email volume, you may also choose to receive your discussion notifications via a daily or weekly digest email as opposed to immediate separate emails.

TM Forum, 240 Headquarters Plaza, East Tower, 10th Floor, Morristown, NJ 07960-6628, USA. Call +1 973 944 5100. info@tmforum.org

[/collapse]

Re: TMF Ontology SIG [#169800]

StephenE6 — Wed, 14 Sep 2011 17:17:34 GMT

Hello, Pierre -

Is the new Ontology SIG community moving forward?

Thanks,

-- Stephen

[collapse]On Aug 23, 2011, at 8:22 PM, Pierre Gauthier wrote:

Pierre Gauthier , CEO and Principal Consultant , Oss Wave

There has been a number of positive responses to this request
"Pierre Gauthier" wrote:
I would like to see if there is enough interest for creating an Ontology SIG within TMF.
I sugest that the initial focus be on establishing the foundation for a TMF Frameworx Ontology.
It is important that the set of proposed ontologies provides complete coverage of the Frameworx. This involves conceptual modeling of eTOM, SID, TAM but also complete coverage of the Business Services and TIP.
Please reply to that post if you are interested in participating to the TMF Ontology SIG.

I will now close this thread and request the creation of the new Ontology SIG community. I recommend you post any Ontology related material on that community. We will pick up the discussion in the new community.

Thanks Pierre
Exploring New Project Ideas
Topics:General

Shared In:4 Communities
Shared In:Information Framework (SID) : General
Business Process Framework (eTOM) : General
TM Forum Integration Program (TIP) : General
Frameworx Architecture Steering Team (FAST) : General
View thread online

Please note, this discussion thread is shared across communities. Any replies that you make to this discussion thread, will be posted to all of the communities and topics listed above.

r>

Full thread available online here:
http://www.tmforum.org/community/groups/information_framework_sid/forum/p/169800/192332.aspx?utm_campaign=community&utm_medium=information_framework
_sid&utm_source=alert#192332

----------------------------------------------------------------------------------------------------------------------------------------------------
This message was sent to dingersoll@versant.com and follows the subscription preferences of your membership of the Information Framework (SID) TM
Forum Online Community.

To unsubscribe immediately, please click here and click the 'Edit Subscriptions' button from the Email Subscription preferences section located on
the right hand side of the page just underneath the menu bar. To cut down on email volume, you may also choose to receive your discussion
notifications via a daily or weekly digest email as opposed to immediate separate emails.

TM Forum, 240 Headquarters Plaza, East Tower, 10th Floor, Morristown, NJ 07960-6628, USA. Call +1 973 944 5100. info@tmforum.org

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 5:54:44 PM
"However, at the current time, it would be premature to suggest that the actual models (and a couple of the transformation engines) would be contributed to the TMF as these component parts currently fall into one of two categories; either proprietary work of a consortium, or proprietary work of a individual/corporation."

Ah!
Would be most interested in TMFs opinion on this as a Forum/Consortium on this point. I agree with Joel, we say no this is not going to happen for various reaons and do other less ambitious but yet useful things as a group, or we say yes we can make this happen and put in the time and effort to come up with work. (New IPR though i understand very little of that document ) it sounds is intended to encourage the latter sort of thing?
Sitaraman

Stephen Estes , Business Process Architect , Cisco Systems
8/16/2011 5:48:11 PM
OWL 2.0 is a current standard, and its subset, OWL-DL is both very expressive and logically tractable. For this, the editing tool, Protege 4.1x, is open source, multi-platform (Java) and robustly maintained.
http://protege.stanford.edu/
-- Stephen

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 5:46:19 PM
And yes, as Ken says one group can feed into the other so the whole becomes greater than the parts with team focused on clearly defined agendas and work in the same environment so that there not avodable software-administrrative hassles that hurt progress. In that sense a Metamodel is indeed an important first step, we need to have our grammar straight before we use it to consistently communicate semantics without chaos due to misinterpretation.
Sitaraman

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 5:42:11 PM
Dave/Jerome,
Yes, I would go with any set of deliverable as long as it is useful to the industry and those are clearly defined.
Having said that, i do feel that in a forum called Ontology SIG, the participants who expressed interest in this group as long as such a group exists should be allowed to go beyond a meta model.
A meta model, to me is the equivalent of an XSD. I helps validation, it checks for the grammar for the language spoken is correct, the statement could very well be to borrow from a poem,
It would be perfectly valid (context awareness, i am reminded of John Strassners post) to say "would like a footlong with everything on it please" in a blockbuster store. So we are back to the killed thread i am afraid, context aware policies and behaviors. But even that belongs in a thread for PBNM or PDNM,
But for a thread that calls itself Ontology SIG should be allowed to explore beyond policies and contexts. even.
Therefore a
a)SID metamodel/Interface Metamodel group can have as a deliverable to come up wth XSD equivalent of Models and Interfaces.for which BTs metamodel doc could be a base. Additional deliverables, automated consistency checks, both static when a new model entities or interfaces are defined(similar to strong typing carrier over to a higher layer here) up to fine grained interface signature/parameter level dynamic checking. We could even extend this to automatic translation between instances of two diferent meta models. About here the name justifies a stop.
b) PBNM which will deal with how to express policies and behaviours an context awareness and determining what are the typical policies that are useful.
c) Onotology SIG, where we go beyond metamodels and context awareness is there a fundametnal structure to knowledge representation, are there fundamental operations on those which would formalize knowledge evolution (collective knowledge, learning). I like Pauls definition, though it could turn out to be ambitious with a narrower focus to our domain by trading off loss in generality, we would have a surer sight of the goal.
Additional deliverables would be,
Exhaustively list concrete Network Semantic Instances.
Map problem space in telecommunication network operations to NEtwork Semantics
Specify using knowledge evolution as above for
   Interface harmonization without loss of semantics
       Simple RCA/SIA
       Self-*
       Trending and Planning.
If the feeling is above is too much to shoot for, i will go with it, but the group should cease to be called Ontology SIG, Ontology and its potential applications to Telecommunications Network
go beyond metamodeling is my feeling.
Sitaraman


Joel Fleck II , Chief Architect , Hewlett-Packard
8/16/2011 5:41:58 PM
Wedge et al:

One of the key directions in research and industry looking forward is the use of a model-driven approach for the construction of business vocabularies, rules, ontologies and technology-neutral representations of the models. In particular, recent research has demonstrated a very usable approach using the OMG’s Semantics for Business Vocabulary and Rules (SBVR) as the basis for transformations to OWL 2.0 (for ontologies and ontological based reasoning –read that semantics), and as the basis for N3 or N-tuple technology-neutral based representations for the policy language(s) of choice, to UML and OCL for the generation of code stubs, and finally to XPATH 2.0 for testing potential interfaces. For reference, I refer to a paper in IEEE Communications August, 2010 “Multidomain IT Architectures for Next-Generation Communications Service Providers”.

Just as an FYI, there has been significant effort on the formulation of an SBVR core for the specification of inter/intra domain relationships and the management of these relationships (and the transformations of that work into UML, OWL, and N3). The methodology has been successfully proven and shown to be very valuable in capturing the business perspective and building software architecture models from them. I would strongly advise investigation into these approaches.

However, at the current time, it would be premature to suggest that the actual models (and a couple of the transformation engines) would be contributed to the TMF as these component parts currently fall into one of two categories; either proprietary work of a consortium, or proprietary work of a individual/corporation.

Best Regards,
Joel

Joel J. Fleck, II
Chief Architect
Head of Standards
HP Software CTO
Hewlett-Packard Company
200 Connell Parkway, Suite 500
Berkeley Heights, NJ 07922

+1 908.898.4091 Phone
+1 908.416.1706 Mobile
+1 908.898.4887 Fax
joel.fleck@hp.com

Kenneth Dilbeck , Director, Technical Program , TM Forum
8/16/2011 5:16:33 PM
Folks,
Anyone familiar with SADL http://sadl.sourceforge.net/?
Perhaps if both groups can use the same tool base, Eclipse, the chances of keeping them together would increase. The JOSIF work is Eclipse based as well.
It would seem having tool support for something as rigorous as an ontology would be helpful. I have only taken a quick look and this may be more oriented towards reasoning systems but thought I would ask.
We do need to make sure there is a very open line of communication between the FAST meta-model team and any other ontology work that takes places to avoid duplication effort and development of conflicting answers.
Regards,
Ken

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 4:47:53 PM
Hi Wedge,
   Isnt modelling itself an abstraction activity, therefore 2 and 1 are same?
Sitaraman

William Greene , Technical SME; CTR DES-E , DOD
8/16/2011 4:46:08 PM
Classification: UNCLASSIFIED
Caveats: NONE

I’m a little shy of using “automated reasoning” as a goal. I’m personally satisfied with “automated services” as a more reachable goal. Perhaps, "automated, adaptive services” as a stretch goal.

I'm simply happy to start with defining the grammar, syntax, semantics. Also, maybe a few basic patterns that support first understanding Frameworx as a model and then basic operations on information contained in the model.

Eventually, (but not soon ;-) we might get to a point where we can use this to, for example, “frame the set of available interface solutions for domain X and iterate across these to determine an optimal decision tree for function L: federate with the similar decision tree coming from Domain Y.

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

David Milham , Chief Architect , TM Forum
8/16/2011 4:30:55 PM
One needed to be on the call to hear the full background and at the presentation at TAW Baltimore
The key point made by Steve was that this was an example of an enterprise repository solution some aspects of which need to be embedded in Frameworx, Without such a metamodel SP like BT assert that is it difficult to practical evolve systems interfaces business services etc. And it is not primarily a technical issue it is a management and governance need on Frameworx.

Dave
Dave Milham
Chief Architect
t +44 5601 855561m+44 7802 902451
edmilham@tmforum.org

From:Exploring New Project Ideas [mailto:newprojects@community.tmforum.org] On Behalf Of Sitaraman Wills
Sent: 16 August 2011 21:14
To: Dave Milham
Subject: [Exploring New Project Ideas] TMF Ontology SIG [#169800]

Sitaraman Wills, Senior Software Engineer , NMSWorks Software Pvt. Ltd.
"David Milham" wrote:
Hi Dave,
Need a favor,

The following document seems to be a MetaModel proposal from Steve Orobeck, this looks like more of a solution document than a list of problems bothering a service provider as you seem to have intended. Am i looking at the right document and if so a pointer to where the problem statements are would help.

If you want to see what is bothering at least one service provider you should look up this contribution.
http://www.tmforum.org/Community/groups/frameworxarchitecture/contributions.aspx?id=artf2509
which was originally presented at TAW Baltimore in the cross Frameworx meetings

The following document gives me a permission denied error, would much aprpeicate if it can be sent as an attachment.
Thanks much.
Sitaraman

http://collab.tmforum.org/integration/viewcvs/viewcvs.cgi/trunk/?root=business_services&system=exsy1002

but we discovered that the individual meta-models don’t line up as we would wish.
Hence we started a top down exercise to complement the prior approach.

What we need now is hand and bodies to work through the patterns we need to use and identify and rectify the gaps in the existing component frameworks.

Dave
Dave Milham
Chief Architect
t +44 5601 855561m+44 7802 902451
edmilham@tmforum.org

From:Information Framework (SID) [mailto:sid@community.tmforum.org] On Behalf Of Sitaraman Wills
Sent: 16 August 2011 19:05
To: Dave Milham
Subject: [Information Framework (SID)] TMF Ontology SIG [#169800]

Sitaraman Wills, Senior Software Engineer , NMSWorks Software Pvt. Ltd.
At the end fo the day, we are TMF first and then Ontologists (to various levels of expertise, from bystanders, through believers, via budding through to been-there-done-thats ).
Perhaps we must have another line of attack from the ServiceProviders side. We could ask operators running their networks day in and day out...tell me in plain english or you native language what are the things that bother you in your job.
There must be complete freedom in responding, no tracking along a direction with a set of 10 questions and the like.
Then we huddle together and extract semantics out of them. Then, we can have that comfortable feeling that, at the end of the day, what we come up with is going to make an important set of people that drive an all pervading industry, smile, courtesy TMF.
Sitaraman
Information Framework (SID)
Topics:General

Shared In:4 Communities
Exploring New Project Ideas
Topics:General

Shared In:4 Communities

Paul Hanke , Sr Professional Staff , Johns Hopkins University Applied Physics Lab
8/16/2011 4:15:48 PM
... I'm not saying that this is what everyone thinks of when they hear the word "ontology", but here is what a deliverable could look like:

An extensible, machine-readable, and foundational logical theory of service provider operations that supports automated reasoning and automated interoperability even across arbitrary domain-specific extensions.

That is, unlike a data model expressed in XML Schema where the best a machine can do with instance data is simply ingore any extensions it is unfamiliar with, an ontology expressed in OWL opens the possibility that a machine can formally manipulate instance data (in a way that mimics human understanding) to include extensions that it has never seen before ... perhaps this is what the FAST team is trying to address with meta-models, I couldn't say ... but even if it is, at this formative stage I see room for both thrust areas to poke and prod at the envelope of the possible.

William Greene , Technical SME; CTR DES-E , DOD
8/16/2011 4:15:42 PM
Classification: UNCLASSIFIED
Caveats: NONE

From Stephen, " Also, the JOSIF open source project has define and implemented a meta-model for interfaces."

Brilliant work!

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

David Milham , Chief Architect , TM Forum
8/16/2011 4:08:59 PM
Steve

Correct thanks for reminding us of the location of that information

Dave
Dave Milham
Chief Architect
t  +44 5601 855561  m+44 7802 902451
edmilham@tmforum.org

From:Exploring New Project Ideas [mailto:newprojects@community.tmforum.org] On Behalf Of Stephen Fratini
Sent: 16 August 2011 20:52
To: Dave Milham
Subject: [Exploring New Project Ideas] TMF Ontology SIG [#169800]

Stephen Fratini, Senior Engineer , Telcordia Technologies
Dave and All,

Also, the JOSIF open source project has define and implemented a meta-model for interfaces.
See the JOSIF wiki pages at
http://sourceforge.net/apps/mediawiki/openoss/index.php?title=Interface_Meta-Model

Steve

---
Stephen Fratini
Telcordia Technologies, Chief Strategy Office
Chair of the TM Forum Integration Program
Home Office: +1 732 962 1500 (I tend to be at this number more often)
Office: +1 732 699 2226
Work Cell: +1 732 853 2207
Email: sfratini@telcordia.com

Exploring New Project Ideas
Topics:General

Shared In:4 Communities

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 4:06:43 PM
"David Milham" wrote:
Hi Dave,
Need a favor,

The following document seems to be a MetaModel proposal from Steve Orobeck, this looks like more of a solution document than a list of problems bothering a service provider as you seem to have intended. Am i looking at the right document and if so a pointer to where the problem statements are would help.

If you want to see what is bothering at least one service provider you should look up this contribution.
http://www.tmforum.org/Community/groups/frameworxarchitecture/contributions.aspx?id=artf2509
which was originally presented at TAW Baltimore in the cross Frameworx meetings

The following document gives me a permission denied error, would much aprpeicate if it can be sent as an attachment.
Thanks much.
Sitaraman

http://collab.tmforum.org/integration/viewcvs/viewcvs.cgi/trunk/?root=business_services&system=exsy1002

but we discovered that the individual meta-models don’t line up as we would wish.
Hence we started a top down exercise to complement the prior approach.

What we need now is hand and bodies to work through the patterns we need to use and identify and rectify the gaps in the existing component frameworks.

Dave
Dave Milham
Chief Architect
t +44 5601 855561m+44 7802 902451
edmilham@tmforum.org

From:Information Framework (SID) [mailto:sid@community.tmforum.org] On Behalf Of Sitaraman Wills
Sent: 16 August 2011 19:05
To: Dave Milham
Subject: [Information Framework (SID)] TMF Ontology SIG [#169800]

Sitaraman Wills, Senior Software Engineer , NMSWorks Software Pvt. Ltd.
At the end fo the day, we are TMF first and then Ontologists (to various levels of expertise, from bystanders, through believers, via budding through to been-there-done-thats ).
Perhaps we must have another line of attack from the ServiceProviders side. We could ask operators running their networks day in and day out...tell me in plain english or you native language what are the things that bother you in your job.
There must be complete freedom in responding, no tracking along a direction with a set of 10 questions and the like.
Then we huddle together and extract semantics out of them. Then, we can have that comfortable feeling that, at the end of the day, what we come up with is going to make an important set of people that drive an all pervading industry, smile, courtesy TMF.
Sitaraman
Information Framework (SID)
Topics:General

Shared In:4 Communities

William Greene , Technical SME; CTR DES-E , DOD
8/16/2011 4:02:35 PM
Classification: UNCLASSIFIED
Caveats: NONE

I think we here and also the FAST team needs to distinguish between meta-model as:
- a model of a model
- an abstraction of a model.

At least in ontology, it should be an abstraction of a model.

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

Stephen Fratini , Senior Engineer , Telcordia Technologies
8/16/2011 3:40:13 PM
Dave and All,

Also, the JOSIF open source project has define and implemented a meta-model for interfaces.
See the JOSIF wiki pages at
http://sourceforge.net/apps/mediawiki/openoss/index.php?title=Interface_Meta-Model

Steve

- Stephen

William Greene , Technical SME; CTR DES-E , DOD
8/16/2011 2:24:04 PM
Classification: UNCLASSIFIED
Caveats: NONE

Yea, Jerome, not so easy ;-)

These are characterizing statements; perhaps true; but neither providing definitions nor establishing relationships.

For example, Behavior and Lifecycle, must be defined in the semantics of the Ontology. The metamodel executes a ontological syntax, a grammar, in applying behavior and Lifecycle to specific targets. (first pass)

Sitaraman, I think, raises an interesting supposition: is the Frameworx metamodel recursive or self referential? Should that be a structural requirement?

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 1:57:06 PM
At the end fo the day, we are TMF first and then Ontologists (to various levels of expertise, from bystanders, through believers, via budding through to been-there-done-thats ).
Perhaps we must have another line of attack from the ServiceProviders side. We could ask operators running their networks day in and day out...tell me in plain english or you native language what are the things that bother you in your job.
There must be complete freedom in responding, no tracking along a direction with a set of 10 questions and the like.
Then we huddle together and extract semantics out of them. Then, we can have that comfortable feeling that, at the end of the day, what we come up with is going to make an important set of people that drive an all pervading industry, smile, courtesy TMF.
Sitaraman

Jerome Hannebelle , Director, Open Innovation, FT/IMG/Technocentre/BP , France Telecom
8/16/2011 1:49:04 PM
Model experts are certainly more qualified than I, but I would think :
1) While an ontology focuses on giving a meaning(= a semantic) to "objects" (sometimes not as formally as a metamodel), a metamodel probably goes beyond in also looking at behavior, lifecycle, ...
2) The metamodel may define abstract, unifying concepts that may not show up in an ontology (because they are abstract).
Mr experts, please jump in !

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 1:34:32 PM
I am particularly interested (from the most generatl interface harmonization perspective) in the following goals.
- Can we structure Semantics (we need not consider the general programming language semantics, a perhaps more difficult problem, but can confine to the most frequently occuring, to begin with, semantic patterns in telecommunication networks) Sub goals would be then
       a) What are the frequently occuring "network semantics" both across and within Networks. We can and perphaps must free ourselves of all constraints and feel free to to be abstract here. Then
       b) Can we represent them in COMMON structural pattern, concretely well defined with as little ambiguity as possible.
- Now that we have a COMMON pattern, for the most popularly occuring (to keep things simple at first) concretizations seen above
        a)Represent different instances of the same concrete Semantics in the common pattern and lay them out.
        b)In the abstract (free form) how do we characterize the "extent of difference" between the two.
        c)How do we concretize the differnce expression (procedures, algorithms on the said common structure) that unambigously characterizes the diff. That is, from the diff, one must be able to get to the other from the other, other and converse.
        d) If we do reach c) then hopefully the holy grail of merge (All of A in B, or All of B in A or B = A+B or B = B-A and other possibilities) will be a simpler extension.
       I am hoping, well i would go the extent of saying it is almost a hunch, that the common strcture would have aspects of self-simlarity. Therefore i am hoping the above operations can be recursively applied to whatever turns out to be the "Atomic Operation" on the said structure and therefore be able to do complex "semantic merges", "Semantic Morphs" or Semantc Overlays and such operations out of the recursive operations composed out of the atomic operations.
      If we reach here we have set ourselves up to casting these a)Structure and b)Unit )perations in telecommunicational semnatic processing (due credits to the Chemical Engineering clan),
    Then if only we get lucky and characterize state (a poorer cousin of full semantics perhaps) with the same pattern, if state is a simpler form of semantics it does seem likely,
-Record State of network entities, of groups of network entities (same state compositional types as above, we can reuse those)
- Give said state(s) as input to Unit Operations and its constituents, and based on the input call suitable child operations.... We have context aware, semantic processing,
- Now have state change as part of the child operation and there you have feedback control, with semantics and syntax, a partial learning system, now we can appeal fo more higher levels of concretized math, Stability, Pole-Zero Analysis and other such intriguing things with which Control Theorists work their systems....?
And here it is... I would much appreciate peoples time in responding why that common structure cannot be the simple TREE of arbitray depth and in general a system represented by a forest of such trees? Concrete examples where it fails in achieving a given functionality would help much
Thanks.
Sitaraman

William Greene , Technical SME; CTR DES-E , DOD
8/16/2011 1:23:22 PM
Classification: UNCLASSIFIED
Caveats: NONE

Jerome, could you expound on " believe relates to FAST team's intent to define the Frameworx's Metamodel." My gut feeling is you are right (that is, I am not disputing this comment at all). I feel this is an important statement; but I need some help seeing exactly how the ontology and the metamodel are linked.

From: Wedge Greene

Craig Gallen , Consultant OpenNMS group , The OpenNMS Group, Inc.
8/16/2011 1:08:15 PM
Hi Pierre,
As discussed, I am very interested in this and would like to include the ten University labs I represent through the web Science trust. www.webscience.org
Cheers
Craig

Jerome Hannebelle , Director, Open Innovation, FT/IMG/Technocentre/BP , France Telecom
8/16/2011 1:06:36 PM
Hi Pierre,

This is definitely an interesting topic, which I believe relates to FAST team's intent to define the Frameworx's Metamodel.
Jerome

William Greene , Technical SME; CTR DES-E , DOD
8/16/2011 1:02:58 PM
Classification: UNCLASSIFIED
Caveats: NONE

I'm interested. However, I think after establishing a basic syntax and some inclusive semantic patterns (e.g., search, aggregation, collaboration, policy), fleshing out the various semantic definitions can be handled in the many traditional workgroups. What we eventually need is for each team to feed their semantics into the overall ontology just like we currently feed the SID and eTOM.

My specific interest today is on how an ontology will work with policy.

I do not see we need to get it perfectly right the first time; just we get it close enough to do much of the work... old 80/20 rule.

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 12:46:06 PM
Hi Pierre,
Meself in it, initially however i request i be granted aceess to discussions and confine myself to observing and learning by participitating in those discussions. I do hope that would be ok.
Thanks.
Sitaraman

Alex Zhdankin , Modeling Architect , Cisco Systems
8/16/2011 12:42:30 PM
Count me in.
Alex.

John Wilmes , Chief Technical Architect, Communications Sector , Progress Software
8/16/2011 12:27:07 PM
I would like to participate especially with reference to SID and eTOM.
John

Paul Hanke , Sr Professional Staff , Johns Hopkins University Applied Physics Lab
8/16/2011 12:27:00 PM
... responding with interest :-)

Bob Natale , Chief Engineer, Joint NetOps & Cyber Defense , MITRE
8/16/2011 12:16:44 PM
Count me in, Pierre.

Pierre Gauthier , CEO and Principal Consultant , Oss Wave
8/16/2011 11:50:18 AM
I would like to see if there is enough interest for creating an Ontology SIG within TMF.
I sugest that the initial focus be on establishing the foundation for a TMF Frameworx Ontology.
It is important that the set of proposed ontologies provides complete coverage of the Frameworx. This involves conceptual modeling of eTOM, SID, TAM but also complete coverage of the Business Services and TIP.
Please reply to that post if you are interested in participating to the TMF Ontology SIG.
End of Discussion History

Full thread available online here: http://www.tmforum.org/community/groups/exploring_new_project_ideas/forum/p/169800/192491.aspx?utm_campaign=community&utm_medium=exploring_new_project_ideas&utm_source=alert#192491
This message was sent to dmilham@tmforum.org and follows the subscription preferences of your membership of the Exploring New Project Ideas TM Forum Online Community.
To unsubscribe immediately, please click here and click the 'Edit Subscriptions' button from the Email Subscription preferences section located on the right hand side of the page just underneath the menu bar. To cut down on email volume, you may also choose to receive your discussion notifications via a daily or weekly digest email as opposed to immediate separate emails.

TM Forum, 240 Headquarters Plaza, East Tower, 10th Floor, Morristown, NJ 07960-6628, USA. Call +1 973 944 5100. info@tmforum.org

This message contains confidential information and is intended only for the individual named. If you are not the named addressee you should not disseminate, distribute or copy this e-mail. Please notify the sender immediately by e-mail if you have received this e-mail by mistake and delete this e-mail from your system. E-mail transmission cannot be guaranteed to be secure or error-free as information could be intercepted, corrupted, lost, destroyed, arrive late or incomplete, or contain viruses. The sender therefore does not accept liability for any errors or omissions in the contents of this message, which arise as a result of e-mail transmission. If verification is required please request a hard-copy version.
Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd. 8/22/2011 12:12:20 AM
Hi,
I think we should just cheat and wait for Bob's approach document, that will likely answer many of these and other questions...
Sitaraman
Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd. 8/21/2011 10:39:11 PM
Hi Bob,
That sounds definitely exciting, From Semantic Web to Semantic Telecommunication Networks....an interesting adventerous and most rewarding journey indeed.
Thanks much for the trigger to take this further.
Sitaraman
Paul Hanke , Sr Professional Staff , Johns Hopkins University Applied Physics Lab 8/21/2011 6:03:34 PM
>>Well, obviously I like this goal - it is relationship-oriented - but why do we need an ontolgy for this? Actually, that's the wrong question, we already have an ontology for this - the UML models, the question should be "why do we need another syntax for this?"

... which provides a nice segue for how the U. S. DoD views "ontology":

"An ontology is used to provide a shared common understanding of a data domain. It is expressed in machine-readable code and captures data domain logic/business rules as data within it. When combined with inference engines, an ontology is used to perform inferencing (reasoning)."

So the DoD thinks we need another syntax in order to achieve automated reasoning about the DoD data universe and all of the information management capabilities that implies (e.g., knowledge discovery, semantic federation, semantic mediation, etc.).

Does that deny that a UML model that is used to generate XML Schema qualifies as "a shared common understanding of a data domain expressed in machine-readable code which captures data domain logic/business rules that when combined with inference engines are used to perform inferencing (reasoning)" ... no ... it just means that the DoD isn't all that fired up about MDA.

Does the TMF need to delve into ontology beyond UML? ... I couldn't say - the OMG work referenced in this thread portrays an architecture that translates UML into OWL, so it would appear that it is possible to obtain the advantages of an ontology technology such as OWL without having to jettison UML (at least up the the limits of expressivity of UML) ... as far as those benefits are concerned, I just read Bob's latest post and he seems to have that base covered ;-) ... i.e., it's not what "ontology" brings to the table that's interesting, but rather the affordances of "ontology technology" that are interesting.

Paul Hanke , Sr Professional Staff , Johns Hopkins University Applied Physics Lab 8/21/2011 4:52:32 PM
>>Another point that needs to eb adressed is the fact that NEs of different hues are still deployed and have their own way ( even if it is a standard way) of structuring their manageable information

... which provides a nice seque for how the U.S. Army envisions ontology technology being applied to help automate interoperability across a diverse community:

Bob Natale , Chief Engineer, Joint NetOps & Cyber Defense , MITRE 8/21/2011 3:36:26 PM
It's clear that as a group we have a long way to go in understanding the value implications of moving to semantic web technologies -- I am confident that the net value is highly positive, after taking the costs into consideration ... but much of the cost is up-front at this stage, and that is usually a hard sell. I think that, in the not very distant future, we (the TM Forum) we align with this value assessment and will aggressively pursue a program -- which I'd like to call Knowledge-Based Operations" -- built around semantic web technologies. As has already been noted, the reasoning power embodied in these technologies (as a set) is at the core, but the value goes way beyond that ... too much to hash out here (usefully), so let me just respond to Francis's comment below, to wit:

" Not to mention tooling - I have given my opinion of OWL/RDF, but if all we have is Protege as tooling, then we are in for some nasty work."

Protégé is not all we have to work with. The W3C's "Semantic Web Development Tools" site -- http://www.w3.org/2001/sw/wiki/Tools -- currently lists 234 tools (79 of which currently work with OWL) -- not all ontology editors or IDEs, of course -- but all from necessary categories for a sufficiently complete solution set.

Of course, SW tools info is just a subset of the info available via the W3C Semantic Web homepage ... I presume most are already familiar with this site, but just in case: http://www.w3.org/2001/sw/wiki/Main_Page.

The set of reasoning operations we want to perform over the Process, Information, Application, and Integration frameworks, among other areas and forms of TM Forum work is way too large to enumerate here ... and is open-ended in terms of growth/refinement anyway, so there is no "final" number of such operations ... a partial list of reasoning operation categories might include:

- Validation
- Normalization
- Harmonization
- Deconfliction
- Optimization
- Rationalization

.. both intra-domain and cross-domain (for any two or more domains, with the "domains" themselves being manifold cross-cutting overlays on our complex operating reality).

Don't worry for now if your potential list of such operations differs from mine, or if you detect overlaps in my categories ... it's just to get the juices flowing ... premature to fix/finalize.

Much other highly useful and valuable work will be done in the TM Forum w/o heavy use of semantic web technologies for several years to come, at least -- I don’t want to discourage anyone from working on those important tasks or to imply any undervaluation of their worth. But we do need a core program driving the move to Knowledge-Based Operations (KBO) -- and I would put planning, discovery, and community education at the top of that group's list for some time, before specification and implementation for example -- sooner rather than later. KBO will tend to highlight:

- Active management (human plane) centered on unity of effort across an agile enterprise
- Operations-centric virtual collaboration environments
- Semantic Web technologies
- Policy-Based Management
- Autonomics

I'll try to come up with a decent diagram illustrating the relationships among those elements in the near future ... it's not just a top-down/bottom-up kind of thing.

Cheers,
BobN

Francis Anderson , BSS/OSS Architect , IBM Corporation 8/21/2011 2:14:43 PM
"Sitaraman Wills" wrote:
To me all the items in goals categorized as external is the intended usage....Would you not consider them so, if so why not?
OK - so let's look at them:
"5. Specification of inter/intra domain relationships[Category: Internal] and the management of these relationships[Category: External]."
Well, obviously I like this goal - it is relationship-oriented - but why do we need an ontolgy for this? Actually, that's the wrong question, we already have an ontology for this - the UML models, the question should be "why do we need another syntax for this?"
"6. Information federation, which is centralto the cross-SDO work on the model and interface federation/harmonization that we've started. (Alex)[Category:External]"
Since when has information federation required an ontology (other than data models), and how would we use the ontology to aid in this? Would this be a separate ontology, crafted for this purpose, or could a general purpose ontology handle it? Wouldn't we have expanded the scope of the ontology from the Frameworx to all SDO Information Models requiring federation?
"7. Shy of using “automated reasoning” , “automated services” as a more reachable goal. Perhaps,"automated, adaptive services” as a stretch goal (Wedge)[Category: External]"
The "shy of" seems to imply that we would not be using the ontology as a basis for reasoning. Without it, wouldn't we just be rendering the Frameworx in "yet another (modeling) syntax"? Is this really what we are trying to do? I don't understand what is meant by "automated services".
Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd. 8/21/2011 1:31:07 PM
Hi Vinay,
This is where i feel we must mix a judicious proportion of epistemology spice in there. The goal of this mix should be to precisely reduce that implementation overhead you are talking of.
If we can
a)Extract out the common semantics that we are dealing with in any network.
b)Extract the hidden structure in these apparently disparate domain representation
c)Relate b) to semantics
d)use a) to correlate them
then we will have a small set of operations with which we can meet the goals that were listed in the summary.
In Materials, Crystalline Materials have only a finite set of possible structures, (Space Groups in Crystal Structure, By Burns and Glazer) Disclaimer: I am not in a position to give answers to penetrating questions from this book, it has gone rusty long time back and i dont own the book..
My hunch is arbitrarily detailed characterizations are mere trees, and knowledge with learning has one or a limited structure to it. If we find that out as Materails Scientists/Matematicians did for their CrystalStructure, suddenly a lot of apparently unrelated, dissimilar things can then be classified (Boomerang....Ontology is helped now ) and handled within a small set of fundametally different units.
Sitaraman
Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd. 8/21/2011 1:14:35 PM
Francis,
I think we hit it there, YES! you have me there,
1. A function y = f(x)
2. A relationship between entity y and entityX
3. And the SC and containedSC(when the change request for this is adopted  is adopted this will come in), read along with the restrictions imposed on the values from the Templates SSC and containedSSC which in turn points to the ServiceDefinition
Are all conceptually rich enough. I dearly hope that this is what you are saying and if so again a big YES we hit it there.
IF the above is true, and this is what you are saying, NOW we can talk about implementation....But let me know if my understandig about what you are saying is correct... then we need to talk more.
Francis Anderson , BSS/OSS Architect , IBM Corporation 8/21/2011 11:46:15 AM
"Sitaraman Wills" wrote:
And here i believe we must move away from entities and relationships, it simply cannot represent a dynamic process which learning is as completely as we like to and we may end up tangling ourselves in profiliferation of classes and relationship which are the baby universes for new set of interface signatures unfortunately.
I would like to throw up a possible paradigm shift, should we start standardizing algorithms?
We're back to the "What is a Relationship?" debate. Personally, and from a mathematical and linguistic perspective, I believe that the relationships should/can/do express the algorithms. From the mathematical point of view, they ARE the function in "y = f(x)", where y is the range and x is the domain. They also are the verb in "subject verb object" where subject is the domain and object is the range.
This is probably the key to our composite SSCV debate - attributes are adjectives, relationships are verbs.
What we then end up with is a Domain Specific Language, with relationships stereotyped much more richly that we tend to in UML. This is where the true patterns are, and when you get it right, there are not that many relationship types, and each one is essentially a variant of the Strategy pattern that wraps the variable (role) by which the relationship is realized in an object.
I have never fully understood your and Alan's assertion that structural modeling is static. To me it is akin to saying language is static.
Vinay Devadatta , Principal Consultant , Wipro Technologies 8/21/2011 11:11:52 AM
Sitaraman,
The issue with a ontology which can adress any type of network (and Manageable instances and types of information ) current and future is that though the interface gets simplified, the implementatio effort at the interface end rises.
Another point that needs to eb adressed is the fact that NEs of different hues are still deployed and have their own way ( even if it is a standard way) of structuring their manageable information
Vinay

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd. 8/21/2011 10:10:15 AM
Hi Francis,
To me all the items in goals categorized as external is the intended usage....Would you not consider them so, if so why not?
I totally identify with your statement
"but if it just comes down to representing existing stuff (semantics) in OWL (syntax), I am not sure that I see the point. Not to mention tooling - I have given my opinion of OWL/RDF, but if all we have is Protege as tooling, then we are in for some nasty work."
I feel the same way.
Wedge is on the dot i feel, that Ontology cannot be viewed and addressed in solitariness, i feel for the likely operator wish list that i had concocted, we need knowledge representation, learning.
We have not so far successfully standardized this, else there wont be all this harmonization, alignment, mapping, amongst various representational entities eating away at our time.
While i completely see with Wedge that there, i differ from Wedge is the means, I believe we can do without a set of policy objects, But it appears that gets into governance, charters and the likes....
And here i believe we must move away from entities and relationships, it simply cannot represent a dynamic process which learning is as completely as we like to and we may end up tangling ourselves in profiliferation of classes and relationship which are the baby universes for new set of interface signatures unfortunately.
I would like to throw up a possible paradigm shift, should we start standardizing algorithms? This is what Steve is proposing for the first phase of operations, we do not have access to the objects, but we will just speficy to the user community how it is done. A first step in standardizing algorithms, why not? That will free us to explore all kinds of learning approaches.
This could be one staging step that i had talked about, which Wedge did not bring into the clear mail thread...?
Sitaraman
Francis Anderson , BSS/OSS Architect , IBM Corporation 8/21/2011 8:32:53 AM
"Pierre Gauthier" wrote:
I sugest that the initial focus be on establishing the foundation for a TMF Frameworx Ontology.
It is important that the set of proposed ontologies provides complete coverage of the Frameworx. This involves conceptual modeling of eTOM, SID, TAM but also complete coverage of the Business Services and TIP.
I must admit that I haven't read all the posts - to quote George (in Yellow Submarine) "It's all too much" - but having been up to my neck in OWL for the past few weeks, I must express interest. To mex mitaphors, after the dust had settled, I thought I might stick my oar in - but I have too many fundamental questions.
What is an ontology? - Is it just OWL?, I certainly hope not - nasty stuff - as if XML weren't bad enough, RDF, and then OWL, please save me!
But there is some sanity out there, Wikipedia, for example:
"In computer science and information science, an ontology formally represents knowledge as a set of concepts within a domain, and the relationships between those concepts. It can be used to reason about the entities within that domain, and may be used to describe the domain."
Isn't that the purpose of the SID, and yet is the SID reasonable?
Another piece of sanity is in the OWL Guide itself:
"The point of this discussion is to note that the development of an ontology should be firmly driven by the intended usage. "
What are we to build the Ontology of - the Frameworx Models, or the meta model - and what is the intended usage? Wedge and Sitaraman did a great job summarizing, but if it just comes down to representing existing stuff (semantics) in OWL (syntax), I am not sure that I see the point. Not to mention tooling - I have given my opinion of OWL/RDF, but if all we have is Protege as tooling, then we are in for some nasty work.
Now, if we were to address the Catalog, that's a different story.
William Greene , Technical SME; CTR DES-E , DOD 8/19/2011 10:16:24 AM
Classification: UNCLASSIFIED
Caveats: NONE

Sorry. POL-3 is just the current SID policy model.

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*
Fidel De la Gala , Manager at IT Strategy and Architecture department , Orange PCS 8/18/2011 8:39:58 PM
Estaré fuera de la oficina hasta el 1 de septiembre. Cualquier tema relacionado con la RFP de Billing convergente poneros en contacto con Alberto Vazquez.

DISCLAIMER:
This message and any attachments (the message) are confidential and intended solely for the addressees. Any unauthorised use or dissemination is prohibited.
Messages are susceptible to alteration. France Telecom Group shall not be liable for the message if altered, changed or falsified.
If you are not the intended addressee of this message, please cancel it immediately and inform the sender.

Este mensaje y sus ficheros adjuntos (el mensaje) son confidenciales y van dirigidos exclusivamente a su/s destinatario/os. Cualquier divulgación o uso no autorizado del mismo queda prohibido.
Los mensajes son susceptibles de alteración. El Grupo France Telecom no se hace responsable del mensaje en caso de alteración, modificación o falsificación del mismo.
Si usted no es el destinatario de este mensaje, por favor elimínelo de forma inmediata y notifíquelo al remitente.
Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd. 8/18/2011 8:11:48 PM
Hi Wedge,
Makes my day when someone such as you think my chipping in has a positive tinge. Thanks
What is POL-3 where i can get information about that?
Thanks
Sitaraman
Shane Furlong , Senior Software Architect , Cable Television Laboratories Inc 8/18/2011 11:52:42 AM
I would be interested as well. We've done some work here at CableLabs already in this area to take the TAM and create an OWL ontology to use in a classification system for application management that might be on interest.
William Greene , Technical SME; CTR DES-E , DOD 8/18/2011 11:30:23 AM
Sitaraman did a good job of summarizing the discussions so far; I have taken the liberty of extracting just that summary into the clear mail stream…. (below from Sitaraman):
“A)Goals:(Items listed in no particular order)
      1. Ontologies provides complete coverage of
the Frameworx. This involves conceptual modeling of eTOM, SID, TAM
but also complete coverage of the Business Services and TIP. (Pierre)[Category: Overall]
      2. An extensible, machine-readable, and foundational logical theory of service provider operations that supports automated reasoning and automated interoperability even across arbitrary domain-specific extensions.(Paul) [Category:Internal]
      3. Meta-models for Frameworx.(Dave)[Category: Internal]
      4. Work through the patterns we need to use and
identify and rectify the gaps in the existing component frameworks.(Dave)[Category: Internal]
      5. Specification of inter/intra domain relationships[Category: Internal] and the management of these relationships[Category: External].
      6. Information federation, which is central
to the cross-SDO work on the model and interface federation/harmonization that we've started. (Alex)[Category:External]
      7. Shy of using “automated reasoning” , “automated services” as a more reachable goal. Perhaps,"automated, adaptive services” as a stretch goal (Wedge)[Category: External]
      8. How an ontology will work with policy.(Wedge)[Category: Internal]
“B)Means(Listed in no particular order):
1.    The use of a model-driven approach for the construction of business vocabularies rules,
ontologies and technology-neutral representations of the models.(Joel)
2.    Formulation of an SBVR core(Joel)
3.    Defining the grammar, syntax, semantics. Also, maybe
a few basic patterns that support first understanding Frameworx
as a model and then basic operations on information contained in the model(Wedge)
4.    What we eventually need is for each team
to feed their semantics into the overall ontology just like we currently feed the SID and eTOM.(Wedge)
"C)Tools (includes languages to express syntactical, ontological, object(class)-relationships, semantics):
OMG’s Semantics for Business Vocabulary and Rules (SBVR) as the basis for transformations to OWL 2.0
(for ontologies and ontological based reasoning –read that semantics)
N3 or N-tuple technology-neutral based representations for the policy
language(s) of choice, to UML and OCL for the generation of code stubs.”
From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*
William Greene , Technical SME; CTR DES-E , DOD 8/18/2011 11:01:58 AM
Classification: UNCLASSIFIED
Caveats: NONE

I am most interested in a 'long-term' work agenda that targets:

- scope of Ontology and policy in Frameworx program: how this relates to metamodel and interface descriptions; can SID and the harmonization work be expressed as an ontology; should we develop ontology as well as UML models?

- standardization on how, in the Frameworx program, the grammar/semantics is expressed; probably following Joel's suggestions on this

- Practical example: standardization of the grammar/structure-of-information for two types of information:

* 'Network or downward: Inventory/configuration/service description/traffic management controls' (Generally, I believe the knowledge engineering should be done in the teams. This specific grammar/structure could then be fed to the TIP program for filling out the branches of the N-tulle)
* description of 'policy' grammar/structure-of-information, that acts on above (frankly, I wanted to do this in the PDNM suggested team. Because I know examples of these exist and hope to tease them out and standardize on an 80/20 amalgam.)

- Development of resources (trained people and tools) that can eventually work to build Frameworx ontology in the various teams. Tie in with the JOSEF work.

I do NOT see this as a team on the Semantic Web. No philosophy. No knowledge (other than SID/eTOM) engineering except as examples.

I believe in the 'short-term', the best we can hope for before February is a technical report on "Ontology as it relates to Frameworx". If this is done in conjunction with PDNM/PDNM, than additionally how it can be used in Policy. I say this because it becomes apparent to me, that if I am to target update of POL-3, the current approaches, as described by Joel, use this avenue to express policy and its targets.

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd. 8/17/2011 6:27:54 PM
Hi,
I thought this would probably be a point, after each of us in the group
expressing individual perspectives, to bring the collective inputs into
one whole. I have attempted to give my understanding at the current point
in time(Please see attached doc)....
Sitaraman
--
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.
Joel Fleck II , Chief Architect , Hewlett-Packard 8/17/2011 12:48:04 PM
Dave:

Absolutely, it would be possible. The IEEE paper and the work in it was based on a keynote that I gave at ManFI 2010. I have also significantly updated the work (the FAME team sponsored and ran the ManFed 2011 workshop at IM 2011 and we did a demo there also). As I wrote in my email, I'm not sure that I can much about the specifics of the inter/intra domain relationship and management models at this time (HP IPR and all), but we sure have a lot of info about the whole process, some of the transformation tools and for that matter an illustrative example (modeling IM 2011 and ManFed 2011 and their relationship - which is an example of inter-domain relationships and potential semantic differences). Guess the main question is the time-frame and the format for the presentation-I just gave an internal seminar on the subject and that covered 1 1/2 hours.

Best regards,
Joel

-- Sent from my HP TouchPad
________________________________
On Aug 17, 2011 11:48, David Milham wrote:

David Milham , Chief Architect , TM Forum
Joel,
This is a very helpful input and the areas you cite seems to be reasonably scoped activities for an Ontology SIG. My concern with the original proposal was that whilst we do need to look at ontology related matters the proposal as stated was very broad, ambitious and will compete for limited SP resources from current date committed activities.

Do you think it would be possible to get a presentation of the IEEE paper from either yourself or Sven?

Dave
Dave Milham
Chief Architect
t +44 5601 855561 m+44 7802 902451
edmilham@tmforum.org

From: Exploring New Project Ideas [mailto:newprojects@community.tmforum.org] On Behalf Of Joel Fleck II
Sent: 16 August 2011 22:49
To: Dave Milham
Subject: [Exploring New Project Ideas] TMF Ontology SIG [#169800]

Joel Fleck II , Chief Architect , Hewlett-Packard

Wedge et al:

One of the key directions in research and industry looking forward is the use of a model-driven approach for the construction of business vocabularies, rules, ontologies and technology-neutral representations of the models. In particular, recent research has demonstrated a very usable approach using the OMG’s Semantics for Business Vocabulary and Rules (SBVR) as the basis for transformations to OWL 2.0 (for ontologies and ontological based reasoning –read that semantics), and as the basis for N3 or N-tuple technology-neutral based representations for the policy language(s) of choice, to UML and OCL for the generation of code stubs, and finally to XPATH 2.0 for testing potential interfaces. For reference, I refer to a paper in IEEE Communications August, 2010 “Multidomain IT Architectures for Next-Generation Communications Service Providers”.

Just as an FYI, there has been significant effort on the formulation of an SBVR core for the specification of inter/intra domain relationships and the management of these relationships (and the transformations of that work into UML, OWL, and N3). The methodology has been successfully proven and shown to be very valuable in capturing the business perspective and building software architecture models from them. I would strongly advise investigation into these approaches.

However, at the current time, it would be premature to suggest that the actual models (and a couple of the transformation engines) would be contributed to the TMF as these component parts currently fall into one of two categories; either proprietary work of a consortium, or proprietary work of a individual/corporation.

Best Regards,
Joel

________________________________

[HP Circle No Bkgrbd]

Joel J. Fleck, II
Chief Architect
Head of Standards
HP Software CTO
Hewlett-Packard Company
200 Connell Parkway, Suite 500
Berkeley Heights, NJ 07922

+1 908.898.4091 Phone
+1 908.416.1706 Mobile
+1 908.898.4887 Fax
joel.fleck@hp.com

________________________________

Exploring New Project Ideas
Topics:General

Shared In:4 Communities

Exploring New Project Ideas
Topics:General

Shared In:4 Communities
________________________________
Shared In:Information Framework (SID) : General
Business Process Framework (eTOM) : General
TM Forum Integration Program (TIP) : General
Frameworx Architecture Steering Team (FAST) : General

View thread online

Please note, this discussion thread is shared across communities. Any replies that you make to this discussion thread, will be posted to all of the communities and topics listed above.

Discussion History
Raghavendra Tadepalli , Architect , Wipro Technologies 8/17/2011 11:01:44 AM

Count me in for this.

Br,

Raghu.

Alex Zhdankin , Modeling Architect , Cisco Systems 8/16/2011 6:37:51 PM

One of the topics the TM Forum Ontology SIG may be focusing on or at least should consider is related to the problem of the information federation, which is central to the cross-SDO work on the model and interface federation/harmonization that we've started.

Some time ago we (TM Forum) has filed through the OMG Telecom PSIG an RFI on Using Problem Specific Views for Information Federation essentially asking for the methodology that would enable creation of the truly connected federated information ecosystem. Apparently the issues described in the RFI resonated quite well with work that been performed by the OMG Architecture Ecosystem SIG (AESIG). The AESIG sent us the response to the RFI referring to the Semantic Information Modeling for Federation (SIMF) RFP, which they've prepared.

Both RFI and the response can be found here:

http://collab.tmforum.org/sf/docman/do/listDocuments/projects.integration_framework/docman.root.tooling_task_force_documents.working_documents_for_end_to_end.omg_rfi

Here is the link to the OMG Wiki where the actual SIMF RFP as well as related discussions can be found:

* http://www.omgwiki.org/architecture-ecosystem/doku.php?id=semantic_information_modeling_for_federation_rfp

Regards,

Alex.

Alex Zhdankin , Modeling Architect , Cisco Systems 8/16/2011 6:23:39 PM

One of the areas where the Ontology SIG may focus on ot at least consider is related to the issue of the information federation, which is central to the cross-SDO work on the model and iterface federation/harmonization.

Some time ago we (TM Forum) filed and RFI with OMG Telecom PSIG asking for the methodology allowing to define the problem-specific views for information federation. This RFI has resonated quite well with the work that already has been started by the OMG Architecture Ecosystem SIG(AESIG), which has sent the Semantic Information Modeling for Federation (SIMF) RFP as a response to our RFI.

You can find both the original RFI and the response here:

http://collab.tmforum.org/sf/docman/do/listDocuments/projects.integration_framework/docman.root.tooling_task_force_documents.working_documents_for_end_to_end.omg_rfi

The text of the SIMF RFP and the related discussions can be found here:

* http://www.omgwiki.org/architecture-ecosystem/doku.php?id=semantic_information_modeling_for_federation_rfp

Regards,

Alex.

David Ingersoll , VP Sales , Versant Corporation (Corporate D) 8/16/2011 6:12:35 PM
All,

Would it be useful for us, Versant, to provide a persistence mechanism for
the ontology. If so, I would be glad to assign a resource to facilitate
working with the group.

David Ingersoll | VP, Sales Americas & APAC
Direct: (650) 232-2430 | EMail: dingersoll@versant.com

(Embedded image moved to file: pic20640.gif) (Embedded image
moved to file: pic26052.gif)
255 Shoreline Drive, Suite 450, Redwood City, CA 94065 | main:
650.232.2400 | fax: 650.232.2401 | www.versant.com | www.db4o.com

CONFIDENTIALITY NOTICE: This e-mail message, including any attachments, is
for the sole use of the intended recipient(s) and may contain confidential
or proprietary information. Any unauthorized review, use, disclosure or
distribution is prohibited. If you are not the intended recipient,
immediately contact the sender by reply e-mail and destroy all copies of
the original message

From: "Joel Fleck II"

To: dingersoll@versant.com

Date: 08/16/2011 02:51 PM

Subject: [Information Framework (SID)] TMF Ontology SIG [#169800]

Sent by: "Information Framework (SID)"

Joel Fleck II , Chief Architect , Hewlett-Packard

Wedge et al:

One of the key directions in research and industry looking forward is the use of a model-driven approach for the construction of business
vocabularies, rules, ontologies and technology-neutral representations of the models. In particular, recent research has demonstrated a very usable
approach using the OMG’s Semantics for Business Vocabulary and Rules (SBVR) as the basis for transformations to OWL 2.0 (for ontologies and
ontological based reasoning –read that semantics), and as the basis for N3 or N-tuple technology-neutral based representations for the policy
language(s) of choice, to UML and OCL for the generation of code stubs, and finally to XPATH 2.0 for testing potential interfaces. For reference, I
refer to a paper in IEEE Communications August, 2010 “Multidomain IT Architectures for Next-Generation Communications Service Providers”.

Just as an FYI, there has been significant effort on the formulation of an SBVR core for the specification of inter/intra domain relationships and
the management of these relationships (and the transformations of that work into UML, OWL, and N3). The methodology has been successfully proven and
shown to be very valuable in capturing the business perspective and building software architecture models from them. I would strongly advise
investigation into these approaches.

However, at the current time, it would be premature to suggest that the actual models (and a couple of the transformation engines) would be
contributed to the TMF as these component parts currently fall into one of two categories; either proprietary work of a consortium, or proprietary
work of a individual/corporation.

Best Regards,

Joel

HP Circle No Bkgrbd
Joel J. Fleck, II
Chief Architect

Head of Standards

HP Software CTO
Hewlett-Packard Company
200 Connell Parkway, Suite 500

Berkeley Heights, NJ 07922

+1 908.898.4091 Phone
+1 908.416.1706 Mobile
+1 908.898.4887 Fax
joel.fleck@hp.com

Information Framework (SID)
Topics:General
Shared In:4 Communities
Shared In:Business Process Framework (eTOM) : General
TM Forum Integration Program (TIP) : General
Frameworx Architecture Steering Team (FAST) : General
Exploring New Project Ideas : General

View thread online

Please note, this discussion thread is shared across communities. Any replies that you make to this discussion thread, will be posted to all of the
communities and topics listed above.

Discussion History

Stephen Estes , Business Process Architect , Cisco Systems
8/16/2011 5:48:11 PM

OWL 2.0 is a current standard, and its subset, OWL-DL is both very expressive and logically tractable. For this, the editing tool, Protege 4.1x, is
open source, multi-platform (Java) and robustly maintained.

http://protege.stanford.edu/

-- Stephen

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 5:46:19 PM

And yes, as Ken says one group can feed into the other so the whole becomes greater than the parts with team focused on clearly defined agendas and
work in the same environment so that there not avodable software-administrrative hassles that hurt progress. In that sense a Metamodel is indeed an
important first step, we need to have our grammar straight before we use it to consistently communicate semantics without chaos due to
misinterpretation.

Sitaraman

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 5:42:11 PM

Dave/Jerome,

Yes, I would go with any set of deliverable as long as it is useful to the industry and those are clearly defined.

Having said that, i do feel that in a forum called Ontology SIG, the participants who expressed interest in this group as long as such a group
exists should be allowed to go beyond a meta model.

A meta model, to me is the equivalent of an XSD. I helps validation, it checks for the grammar for the language spoken is correct, the statement
could very well be to borrow from a poem,

It would be perfectly valid (context awareness, i am reminded of John Strassners post) to say "would like a footlong with everything on it
please" in a blockbuster store. So we are back to the killed thread i am afraid, context aware policies and behaviors. But even that belongs in a
thread for PBNM or PDNM,

But for a thread that calls itself Ontology SIG should be allowed to explore beyond policies and contexts. even.

Therefore a

a)SID metamodel/Interface Metamodel group can have as a deliverable to come up wth XSD equivalent of Models and Interfaces.for which BTs metamodel
doc could be a base. Additional deliverables, automated consistency checks, both static when a new model entities or interfaces are defined(similar
to strong typing carrier over to a higher layer here) up to fine grained interface signature/parameter level dynamic checking. We could even extend
this to automatic translation between instances of two diferent meta models. About here the name justifies a stop.

b) PBNM which will deal with how to express policies and behaviours an context awareness and determining what are the typical policies that are
useful.

c) Onotology SIG, where we go beyond metamodels and context awareness is there a fundametnal structure to knowledge representation, are there
fundamental operations on those which would formalize knowledge evolution (collective knowledge, learning). I like Pauls definition, though it could
turn out to be ambitious with a narrower focus to our domain by trading off loss in generality, we would have a surer sight of the goal.

Additional deliverables would be,

Exhaustively list concrete Network Semantic Instances.

Map problem space in telecommunication network operations to NEtwork Semantics

Specify using knowledge evolution as above for

Interface harmonization without loss of semantics

Simple RCA/SIA

Self-*

Trending and Planning.

If the feeling is above is too much to shoot for, i will go with it, but the group should cease to be called Ontology SIG, Ontology and its
potential applications to Telecommunications Network
go beyond metamodeling is my feeling.

Sitaraman

Kenneth Dilbeck , Director, Technical Program , TM Forum
8/16/2011 5:16:33 PM

Folks,

Anyone familiar with SADL http://sadl.sourceforge.net/?

Perhaps if both groups can use the same tool base, Eclipse, the chances of keeping them together would increase. The JOSIF work is Eclipse based as
well.

It would seem having tool support for something as rigorous as an ontology would be helpful. I have only taken a quick look and this may be more
oriented towards reasoning systems but thought I would ask.

We do need to make sure there is a very open line of communication between the FAST meta-model team and any other ontology work that takes places to
avoid duplication effort and development of conflicting answers.

Regards,

Ken

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 4:47:53 PM

Hi Wedge,

Isnt modelling itself an abstraction activity, therefore 2 and 1 are same?

Sitaraman

William Greene , Technical SME; CTR DES-E , DOD
8/16/2011 4:46:08 PM

Classification: UNCLASSIFIED
Caveats: NONE

I’m a little shy of using “automated reasoning” as a goal. I’m personally satisfied with “automated services” as a more reachable goal. Perhaps,
"automated, adaptive services” as a stretch goal.

I'm simply happy to start with defining the grammar, syntax, semantics. Also, maybe a few basic patterns that support first understanding Frameworx
as a model and then basic operations on information contained in the model.

Eventually, (but not soon ;-) we might get to a point where we can use this to, for example, “frame the set of available interface solutions for
domain X and iterate across these to determine an optimal decision tree for function L: federate with the similar decision tree coming from Domain
Y.

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

David Milham , Chief Architect , TM Forum
8/16/2011 4:30:55 PM

One needed to be on the call to hear the full background and at the presentation at TAW Baltimore

The key point made by Steve was that this was an example of an enterprise repository solution some aspects of which need to be embedded in
Frameworx, Without such a metamodel SP like BT assert that is it difficult to practical evolve systems interfaces business services etc. And it is
not primarily a technical issue it is a management and governance need on Frameworx.

Dave

Dave Milham

Chief Architect

t +44 5601 855561 m+44 7802 902451
edmilham@tmforum.org

From: Exploring New Project Ideas [mailto:newprojects@community.tmforum.org] On Behalf Of Sitaraman Wills
Sent: 16 August 2011 21:14
To: Dave Milham
Subject: [Exploring New Project Ideas] TMF Ontology SIG [#169800]

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.

"David Milham" wrote:

Hi Dave,

Need a favor,

The following document seems to be a MetaModel proposal from Steve Orobeck, this looks like more of a solution document than a list of problems
bothering a service provider as you seem to have intended. Am i looking at the right document and if so a pointer to where the problem
statements are would help.

If you want to see what is bothering at least one service provider you should look up this contribution.

http://www.tmforum.org/Community/groups/frameworxarchitecture/contributions.aspx?id=artf2509

which was originally presented at TAW Baltimore in the cross Frameworx meetings

The following document gives me a permission denied error, would much aprpeicate if it can be sent as an attachment.

Thanks much.

Sitaraman

http://collab.tmforum.org/integration/viewcvs/viewcvs.cgi/trunk/?root=business_services&system=exsy1002

but we discovered that the individual meta-models don’t line up as we would wish.

Hence we started a top down exercise to complement the prior approach.

What we need now is hand and bodies to work through the patterns we need to use and identify and rectify the gaps in the existing component
frameworks.

Dave

Dave Milham

Chief Architect

t +44 5601 855561m+44 7802 902451
edmilham@tmforum.org

From:Information Framework (SID) [mailto:sid@community.tmforum.org] On Behalf Of Sitaraman Wills
Sent: 16 August 2011 19:05
To: Dave Milham
Subject: [Information Framework (SID)] TMF Ontology SIG [#169800]

Sitaraman Wills, Senior Software Engineer , NMSWorks Software Pvt. Ltd.

At the end fo the day, we are TMF first and then Ontologists (to various levels of expertise, from bystanders, through believers, via budding
through to been-there-done-thats ).

Perhaps we must have another line of attack from the ServiceProviders side. We could ask operators running their networks day in and day out...tell
me in plain english or you native language what are the things that bother you in your job.

There must be complete freedom in responding, no tracking along a direction with a set of 10 questions and the like.

Then we huddle together and extract semantics out of them. Then, we can have that comfortable feeling that, at the end of the day, what we come up
with is going to make an important set of people that drive an all pervading industry, smile, courtesy TMF.

Sitaraman
Information Framework (SID)
Topics:General

Shared In:4 Communities

Exploring New Project Ideas
Topics:General

Shared In:4 Communities

Paul Hanke , Sr Professional Staff , Johns Hopkins University Applied Physics Lab
8/16/2011 4:15:48 PM

. I'm not saying that this is what everyone thinks of when they hear the word "ontology", but here is what a deliverable could look like:

An extensible, machine-readable, and foundational logical theory of service provider operations that supports automated reasoning and automated
interoperability even across arbitrary domain-specific extensions.

That is, unlike a data model expressed in XML Schema where the best a machine can do with instance data is simply ingore any extensions it is
unfamiliar with, an ontology expressed in OWL opens the possibility that a machine can formally manipulate instance data (in a way that mimics human
understanding) to include extensions that it has never seen before . perhaps this is what the FAST team is trying to address with meta-models, I
couldn't say ... but even if it is, at this formative stage I see room for both thrust areas to poke and prod at the envelope of the possible.

William Greene , Technical SME; CTR DES-E , DOD
8/16/2011 4:15:42 PM

Classification: UNCLASSIFIED
Caveats: NONE

From Stephen, " Also, the JOSIF open source project has define and implemented a meta-model for interfaces."

Brilliant work!

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

David Milham , Chief Architect , TM Forum
8/16/2011 4:08:59 PM

Steve

Correct thanks for reminding us of the location of that information

Dave

Dave Milham

Chief Architect

t +44 5601 855561 m+44 7802 902451
edmilham@tmforum.org

From: Exploring New Project Ideas [mailto:newprojects@community.tmforum.org] On Behalf Of Stephen Fratini
Sent: 16 August 2011 20:52
To: Dave Milham
Subject: [Exploring New Project Ideas] TMF Ontology SIG [#169800]

Stephen Fratini , Senior Engineer , Telcordia Technologies

Dave and All,

Also, the JOSIF open source project has define and implemented a meta-model for interfaces.

See the JOSIF wiki pages at

http://sourceforge.net/apps/mediawiki/openoss/index.php?title=Interface_Meta-Model

Steve

---

Stephen Fratini

Telcordia Technologies, Chief Strategy Office

Chair of the TM Forum Integration Program

Home Office: +1 732 962 1500 (I tend to be at this number more often)

Office: +1 732 699 2226

Work Cell: +1 732 853 2207

Email: sfratini@telcordia.com

Exploring New Project Ideas
Topics:General

Shared In:4 Communities

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 4:06:43 PM

"David Milham" wrote:

Hi Dave,

Need a favor,

The following document seems to be a MetaModel proposal from Steve Orobeck, this looks like more of a solution document than a list of problems
bothering a service provider as you seem to have intended. Am i looking at the right document and if so a pointer to where the problem
statements are would help.

If you want to see what is bothering at least one service provider you should look up this contribution.

http://www.tmforum.org/Community/groups/frameworxarchitecture/contributions.aspx?id=artf2509

which was originally presented at TAW Baltimore in the cross Frameworx meetings

The following document gives me a permission denied error, would much aprpeicate if it can be sent as an attachment.

Thanks much.

Sitaraman

http://collab.tmforum.org/integration/viewcvs/viewcvs.cgi/trunk/?root=business_services&system=exsy1002

but we discovered that the individual meta-models don’t line up as we would wish.

Hence we started a top down exercise to complement the prior approach.

What we need now is hand and bodies to work through the patterns we need to use and identify and rectify the gaps in the existing component
frameworks.

Dave

Dave Milham

Chief Architect

t +44 5601 855561m+44 7802 902451
edmilham@tmforum.org

From:Information Framework (SID) [mailto:sid@community.tmforum.org] On Behalf Of Sitaraman Wills
Sent: 16 August 2011 19:05
To: Dave Milham
Subject: [Information Framework (SID)] TMF Ontology SIG [#169800]

Sitaraman Wills, Senior Software Engineer , NMSWorks Software Pvt. Ltd.

At the end fo the day, we are TMF first and then Ontologists (to various levels of expertise, from bystanders, through believers, via budding
through to been-there-done-thats ).

Perhaps we must have another line of attack from the ServiceProviders side. We could ask operators running their networks day in and day out...tell
me in plain english or you native language what are the things that bother you in your job.

There must be complete freedom in responding, no tracking along a direction with a set of 10 questions and the like.

Then we huddle together and extract semantics out of them. Then, we can have that comfortable feeling that, at the end of the day, what we come up
with is going to make an important set of people that drive an all pervading industry, smile, courtesy TMF.

Sitaraman
Information Framework (SID)
Topics:General

Shared In:4 Communities

William Greene , Technical SME; CTR DES-E , DOD
8/16/2011 4:02:35 PM

Classification: UNCLASSIFIED
Caveats: NONE

I think we here and also the FAST team needs to distinguish between meta-model as:
- a model of a model
- an abstraction of a model.

At least in ontology, it should be an abstraction of a model.

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

Stephen Fratini , Senior Engineer , Telcordia Technologies
8/16/2011 3:40:13 PM

Dave and All,

Also, the JOSIF open source project has define and implemented a meta-model for interfaces.
See the JOSIF wiki pages at
http://sourceforge.net/apps/mediawiki/openoss/index.php?title=Interface_Meta-Model

Steve

---
Stephen Fratini
Telcordia Technologies, Chief Strategy Office
Chair of the TM Forum Integration Program
Home Office: +1 732 962 1500 (I tend to be at this number more often)
Office: +1 732 699 2226
Work Cell: +1 732 853 2207
Email: sfratini@telcordia.com

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 3:34:26 PM

Hi Dave,

Thanks so much for those concrete pointers, I am already committed to the interface aspects of TMF, but at any open opportunitues i find, here is
an intact body with appendages to boot which will do its best to chip in.

My first priority in this thread will now be to look thro the docs you have pointed out. Timeout from me for now.

Thanks again.

Sitaraman

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 3:29:18 PM

Hi Jerome,

I have unilaterally decided to go against your call, in spite of being a relative novice, i am jumping in. What are your thoughts on the following,
any pointers also would help.

A model is a reflection of something fundamental in the system under study. Those fundamental truths we try to capture in the model.

However whenwe talk of behavior and lifecycle there is an element of "free will" here, it becomes more difficult to assign something too
fundamental to it and if you can it probably belongs as a characteristic. So (i do realize process models are widely accepted now but still if i ask
myself do i truly believe in such a thing i am not able to confidently say yes i do,)process models are too dignfied a name, they are more of
compilations fo similar process instances than a dignified models that abstractts a reality.

To take the line of thought, what Wedge was pointing out, what would be the general opinion on seeing

Information as modelled

States as Characteristic each of them forming/conforming to an integral part of that model.

And behaviour as a function action on a domain set of states and range same set of states

And life cycle as an instance of a compostie/complex behavior chain, a chain of functions, i guess Mathematicians term this as functional
composition...

Mathematical Trees (Graphs including hierarchical bidirectional graphs), Sets, Functions and functional analysis seem to me to one set of
candidate fundamental constructs we could look at...Wedge what would your thoughs be too on this?

Sitaraman

David Milham , Chief Architect , TM Forum
8/16/2011 3:24:56 PM

For information

The FAST team is preparing a second pass at Meta-models for Frameworx.

The TR structure is at

http://collab.tmforum.org/sf/docman/do/listDocuments/projects.integration_framework/docman.root.metamodels.tr1xx

And the first contribution from BT based on an operational enterprise repository was made earlier this week and discussed on today’s FAST calls.

If you want to see what is bothering at least one service provider you should look up this contribution.

http://www.tmforum.org/Community/groups/frameworxarchitecture/contributions.aspx?id=artf2509

which was originally presented at TAW Baltimore in the cross Frameworx meetings

The first attempt was to capture the meta-models in the existing component frameworks . This bottom up approach has been competed using ECORE
metamodelling language an dis on subversion at:

http://collab.tmforum.org/integration/viewcvs/viewcvs.cgi/trunk/?root=business_services&system=exsy1002

but we discovered that the individual meta-models don’t line up as we would wish.

Hence we started a top down exercise to complement the prior approach.

What we need now is hand and bodies to work through the patterns we need to use and identify and rectify the gaps in the existing component
frameworks.

Dave

Dave Milham

Chief Architect

t +44 5601 855561 m+44 7802 902451
edmilham@tmforum.org

From: Information Framework (SID) [mailto:sid@community.tmforum.org] On Behalf Of Sitaraman Wills
Sent: 16 August 2011 19:05
To: Dave Milham
Subject: [Information Framework (SID)] TMF Ontology SIG [#169800]

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.

At the end fo the day, we are TMF first and then Ontologists (to various levels of expertise, from bystanders, through believers, via budding
through to been-there-done-thats ).

Perhaps we must have another line of attack from the ServiceProviders side. We could ask operators running their networks day in and day out...tell
me in plain english or you native language what are the things that bother you in your job.

There must be complete freedom in responding, no tracking along a direction with a set of 10 questions and the like.

Then we huddle together and extract semantics out of them. Then, we can have that comfortable feeling that, at the end of the day, what we come up
with is going to make an important set of people that drive an all pervading industry, smile, courtesy TMF.

Sitaraman
Information Framework (SID)
Topics:General

Shared In:4 Communities

David Milham , Chief Architect , TM Forum
8/16/2011 3:04:40 PM

Jerome

That’s my opinion too. I like to understand better exactly what a deliverable would look like

Dave

Dave Milham

Chief Architect

t +44 5601 855561 m+44 7802 902451
edmilham@tmforum.org

From: Information Framework (SID) [mailto:sid@community.tmforum.org]
Sent: 16 August 2011 18:10
To: Dave Milham
Subject: [Information Framework (SID)] TMF Ontology SIG [#169800]

Jerome Hannebelle , Director, Open Innovation, FT/IMG/Technocentre/BP , France Telecom

Hi Pierre,

This is definitely an interesting topic, which I believe relates to FAST team's intent to define the Frameworx's Metamodel.

Jerome
Information Framework (SID)
Topics:General

Shared In:4 Communities

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 2:54:34 PM

Ok, I hear you Wedge, what in your opinion would be those foundational structures we need to focus on,. I am shooting off the cuff there admitedly,
any pointers to formal references would help. Now that i have had the benefit of free thiking on this subject without any set notions, thanks to you
all i am ready to look at what has been formalized...

Sitaraman

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 2:51:45 PM

Hi Wedge,

Glad you find the thought interesting... before putting on it a rather binding-sounding term such as requirement, how about we will be open, no
tilt either way and see if when the chips fall in place self-similarity shows up. How many ever times we say it, we need a formal mathematic proof
to associate strong property of a requirement, (those "necessary and sufficients" and the likes these mathematicians never find sufficient to use)
that takes us into the relam of mathematics of mathematics truly self similar.

But as a lay man, Knowledge is something you can only get more effectively out of existing knoweldge, that is why as one progresses up the grades,
one learns more using the tools of the existing knowledge, (let us forget for the moment the bing bang of the first baby cry conveniently) self
similarity seems very likely.

Coming to think of it, it appears we do a breadth first, and as you go down the knowledge tree the no of child increases and specialist spend a lot
of timein the numerous number of fine grained children of their parent field and lifetimes are not enough to go to the next sibling of the parent.

As an aside for the educationists then, what woud happen if people were given the choice for part depth-first education?! What would depth first
and breadth first mean in the network semantics mapping we would come up with? For RCA with a high level fault at the top levels of the tree, we
would likely need a BFS, for SIA it is DFS - Duality showing up again!

Sitaraman

William Greene , Technical SME; CTR DES-E , DOD
8/16/2011 2:34:17 PM

Classification: UNCLASSIFIED
Caveats: NONE

Sitaraman, I think I get the gist of this. But I feel you are jumping ahead here. Running too far down the inference road. While I agree, an
ontology which opens the door for your conclusions also would be my goal (graph tree, semantic processing, parent/child service structures...); we
are not there yet. And must rigorously work out the foundation to get there.

But fair disclosure: I am not an expert in ontology.

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

Stephen Estes , Business Process Architect , Cisco Systems
8/16/2011 2:28:34 PM

Yes, I'd be interested.

Currently working with NIST on their cloud ref arch ontology.

-- Stephen

William Greene , Technical SME; CTR DES-E , DOD
8/16/2011 2:24:04 PM

Classification: UNCLASSIFIED
Caveats: NONE

Yea, Jerome, not so easy ;-)

These are characterizing statements; perhaps true; but neither providing definitions nor establishing relationships.

For example, Behavior and Lifecycle, must be defined in the semantics of the Ontology. The metamodel executes a ontological syntax, a grammar, in
applying behavior and Lifecycle to specific targets. (first pass)

Sitaraman, I think, raises an interesting supposition: is the Frameworx metamodel recursive or self referential? Should that be a structural
requirement?

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 1:57:06 PM

At the end fo the day, we are TMF first and then Ontologists (to various levels of expertise, from bystanders, through believers, via budding
through to been-there-done-thats ).

Perhaps we must have another line of attack from the ServiceProviders side. We could ask operators running their networks day in and day out...tell
me in plain english or you native language what are the things that bother you in your job.

There must be complete freedom in responding, no tracking along a direction with a set of 10 questions and the like.

Then we huddle together and extract semantics out of them. Then, we can have that comfortable feeling that, at the end of the day, what we come up
with is going to make an important set of people that drive an all pervading industry, smile, courtesy TMF.

Sitaraman

Jerome Hannebelle , Director, Open Innovation, FT/IMG/Technocentre/BP , France Telecom
8/16/2011 1:49:04 PM

Model experts are certainly more qualified than I, but I would think :

1) While an ontology focuses on giving a meaning(= a semantic) to "objects" (sometimes not as formally as a metamodel), a metamodel probably goes
beyond in also looking at behavior, lifecycle, ...

2) The metamodel may define abstract, unifying concepts that may not show up in an ontology (because they are abstract).

Mr experts, please jump in !

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 1:34:32 PM

I am particularly interested (from the most generatl interface harmonization perspective) in the following goals.

- Can we structure Semantics (we need not consider the general programming language semantics, a perhaps more difficult problem, but can confine
to the most frequently occuring, to begin with, semantic patterns in telecommunication networks) Sub goals would be then

a) What are the frequently occuring "network semantics" both across and within Networks. We can and perphaps must free ourselves of all
constraints and feel free to to be abstract here. Then

b) Can we represent them in COMMON structural pattern, concretely well defined with as little ambiguity as possible.

- Now that we have a COMMON pattern, for the most popularly occuring (to keep things simple at first) concretizations seen above

a)Represent different instances of the same concrete Semantics in the common pattern and lay them out.

b)In the abstract (free form) how do we characterize the "extent of difference" between the two.

c)How do we concretize the differnce expression (procedures, algorithms on the said common structure) that unambigously characterizes the
diff. That is, from the diff, one must be able to get to the other from the other, other and converse.

d) If we do reach c) then hopefully the holy grail of merge (All of A in B, or All of B in A or B = A+B or B = B-A and other possibilities)
will be a simpler extension.

I am hoping, well i would go the extent of saying it is almost a hunch, that the common strcture would have aspects of self-simlarity.
Therefore i am hoping the above operations can be recursively applied to whatever turns out to be the "Atomic Operation" on the said structure and
therefore be able to do complex "semantic merges", "Semantic Morphs" or Semantc Overlays and such operations out of the recursive operations
composed out of the atomic operations.

If we reach here we have set ourselves up to casting these a)Structure and b)Unit )perations in telecommunicational semnatic processing (due
credits to the Chemical Engineering clan),

Then if only we get lucky and characterize state (a poorer cousin of full semantics perhaps) with the same pattern, if state is a simpler form
of semantics it does seem likely,

-Record State of network entities, of groups of network entities (same state compositional types as above, we can reuse those)

- Give said state(s) as input to Unit Operations and its constituents, and based on the input call suitable child operations.... We have context
aware, semantic processing,

- Now have state change as part of the child operation and there you have feedback control, with semantics and syntax, a partial learning system,
now we can appeal fo more higher levels of concretized math, Stability, Pole-Zero Analysis and other such intriguing things with which Control
Theorists work their systems....?

And here it is... I would much appreciate peoples time in responding why that common structure cannot be the simple TREE of arbitray depth and in
general a system represented by a forest of such trees? Concrete examples where it fails in achieving a given functionality would help much

Thanks.

Sitaraman

William Greene , Technical SME; CTR DES-E , DOD
8/16/2011 1:23:22 PM

Classification: UNCLASSIFIED
Caveats: NONE

Jerome, could you expound on " believe relates to FAST team's intent to define the Frameworx's Metamodel." My gut feeling is you are right (that is,
I am not disputing this comment at all). I feel this is an important statement; but I need some help seeing exactly how the ontology and the
metamodel are linked.

From: Wedge Greene

Craig Gallen , Consultant OpenNMS group , The OpenNMS Group, Inc.
8/16/2011 1:08:15 PM

Hi Pierre,

As discussed, I am very interested in this and would like to include the ten University labs I represent through the web Science trust.
www.webscience.org

Cheers

Craig

Jerome Hannebelle , Director, Open Innovation, FT/IMG/Technocentre/BP , France Telecom
8/16/2011 1:06:36 PM

Hi Pierre,

This is definitely an interesting topic, which I believe relates to FAST team's intent to define the Frameworx's Metamodel.

Jerome

William Greene , Technical SME; CTR DES-E , DOD
8/16/2011 1:02:58 PM

Classification: UNCLASSIFIED
Caveats: NONE

I'm interested. However, I think after establishing a basic syntax and some inclusive semantic patterns (e.g., search, aggregation, collaboration,
policy), fleshing out the various semantic definitions can be handled in the many traditional workgroups. What we eventually need is for each team
to feed their semantics into the overall ontology just like we currently feed the SID and eTOM.

My specific interest today is on how an ontology will work with policy.

I do not see we need to get it perfectly right the first time; just we get it close enough to do much of the work... old 80/20 rule.

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 12:46:06 PM

Hi Pierre,

Meself in it, initially however i request i be granted aceess to discussions and confine myself to observing and learning by participitating in
those discussions. I do hope that would be ok.

Thanks.

Sitaraman

Alex Zhdankin , Modeling Architect , Cisco Systems
8/16/2011 12:42:30 PM

Count me in.

Alex.

John Wilmes , Chief Technical Architect, Communications Sector , Progress Software
8/16/2011 12:27:07 PM

I would like to participate especially with reference to SID and eTOM.

John

Paul Hanke , Sr Professional Staff , Johns Hopkins University Applied Physics Lab
8/16/2011 12:27:00 PM

. responding with interest :-)

Bob Natale , Chief Engineer, Joint NetOps & Cyber Defense , MITRE
8/16/2011 12:16:44 PM

Count me in, Pierre.

Pierre Gauthier , CEO and Principal Consultant , Oss Wave
8/16/2011 11:50:18 AM

I would like to see if there is enough interest for creating an Ontology SIG within TMF.

I sugest that the initial focus be on establishing the foundation for a TMF Frameworx Ontology.

It is important that the set of proposed ontologies provides complete coverage of the Frameworx. This involves conceptual modeling of eTOM, SID, TAM
but also complete coverage of the Business Services and TIP.

Please reply to that post if you are interested in participating to the TMF Ontology SIG.

End of Discussion History

Full thread available online here:
http://www.tmforum.org/community/groups/information_framework_sid/forum/p/169800/192332.aspx?utm_campaign=community&utm_medium=information_framework
_sid&utm_source=alert#192332

----------------------------------------------------------------------------------------------------------------------------------------------------
This message was sent to dingersoll@versant.com and follows the subscription preferences of your membership of the Information Framework (SID) TM
Forum Online Community.

To unsubscribe immediately, please click here and click the 'Edit Subscriptions' button from the Email Subscription preferences section located on
the right hand side of the page just underneath the menu bar. To cut down on email volume, you may also choose to receive your discussion
notifications via a daily or weekly digest email as opposed to immediate separate emails.

TM Forum, 240 Headquarters Plaza, East Tower, 10th Floor, Morristown, NJ 07960-6628, USA. Call +1 973 944 5100. info@tmforum.org

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd. 8/16/2011 5:54:44 PM

"However, at the current time, it would be premature to suggest that the actual models (and a couple of the transformation engines) would be contributed to the TMF as these component parts currently fall into one of two categories; either proprietary work of a consortium, or proprietary work of a individual/corporation."

Ah!

Would be most interested in TMFs opinion on this as a Forum/Consortium on this point. I agree with Joel, we say no this is not going to happen for various reaons and do other less ambitious but yet useful things as a group, or we say yes we can make this happen and put in the time and effort to come up with work. (New IPR though i understand very little of that document ) it sounds is intended to encourage the latter sort of thing?

Sitaraman

Stephen Estes , Business Process Architect , Cisco Systems 8/16/2011 5:48:11 PM

OWL 2.0 is a current standard, and its subset, OWL-DL is both very expressive and logically tractable. For this, the editing tool, Protege 4.1x, is open source, multi-platform (Java) and robustly maintained.

http://protege.stanford.edu/

-- Stephen

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd. 8/16/2011 5:46:19 PM

And yes, as Ken says one group can feed into the other so the whole becomes greater than the parts with team focused on clearly defined agendas and work in the same environment so that there not avodable software-administrrative hassles that hurt progress. In that sense a Metamodel is indeed an important first step, we need to have our grammar straight before we use it to consistently communicate semantics without chaos due to misinterpretation.

Sitaraman

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd. 8/16/2011 5:42:11 PM

Dave/Jerome,

Yes, I would go with any set of deliverable as long as it is useful to the industry and those are clearly defined.

Having said that, i do feel that in a forum called Ontology SIG, the participants who expressed interest in this group as long as such a group exists should be allowed to go beyond a meta model.

A meta model, to me is the equivalent of an XSD. I helps validation, it checks for the grammar for the language spoken is correct, the statement could very well be to borrow from a poem,

It would be perfectly valid (context awareness, i am reminded of John Strassners post) to say "would like a footlong with everything on it please" in a blockbuster store. So we are back to the killed thread i am afraid, context aware policies and behaviors. But even that belongs in a thread for PBNM or PDNM,

But for a thread that calls itself Ontology SIG should be allowed to explore beyond policies and contexts. even.

Therefore a

a)SID metamodel/Interface Metamodel group can have as a deliverable to come up wth XSD equivalent of Models and Interfaces.for which BTs metamodel doc could be a base. Additional deliverables, automated consistency checks, both static when a new model entities or interfaces are defined(similar to strong typing carrier over to a higher layer here) up to fine grained interface signature/parameter level dynamic checking. We could even extend this to automatic translation between instances of two diferent meta models. About here the name justifies a stop.

b) PBNM which will deal with how to express policies and behaviours an context awareness and determining what are the typical policies that are useful.

c) Onotology SIG, where we go beyond metamodels and context awareness is there a fundametnal structure to knowledge representation, are there fundamental operations on those which would formalize knowledge evolution (collective knowledge, learning). I like Pauls definition, though it could turn out to be ambitious with a narrower focus to our domain by trading off loss in generality, we would have a surer sight of the goal.

Additional deliverables would be,

Exhaustively list concrete Network Semantic Instances.

Map problem space in telecommunication network operations to NEtwork Semantics

Specify using knowledge evolution as above for

Interface harmonization without loss of semantics

Simple RCA/SIA

Self-*

Trending and Planning.

If the feeling is above is too much to shoot for, i will go with it, but the group should cease to be called Ontology SIG, Ontology and its potential applications to Telecommunications Network
go beyond metamodeling is my feeling.

Sitaraman

Joel Fleck II , Chief Architect , Hewlett-Packard 8/16/2011 5:41:58 PM
Wedge et al:

One of the key directions in research and industry looking forward is the use of a model-driven approach for the construction of business vocabularies, rules, ontologies and technology-neutral representations of the models. In particular, recent research has demonstrated a very usable approach using the OMG’s Semantics for Business Vocabulary and Rules (SBVR) as the basis for transformations to OWL 2.0 (for ontologies and ontological based reasoning –read that semantics), and as the basis for N3 or N-tuple technology-neutral based representations for the policy language(s) of choice, to UML and OCL for the generation of code stubs, and finally to XPATH 2.0 for testing potential interfaces. For reference, I refer to a paper in IEEE Communications August, 2010 “Multidomain IT Architectures for Next-Generation Communications Service Providers”.

Just as an FYI, there has been significant effort on the formulation of an SBVR core for the specification of inter/intra domain relationships and the management of these relationships (and the transformations of that work into UML, OWL, and N3). The methodology has been successfully proven and shown to be very valuable in capturing the business perspective and building software architecture models from them. I would strongly advise investigation into these approaches.

However, at the current time, it would be premature to suggest that the actual models (and a couple of the transformation engines) would be contributed to the TMF as these component parts currently fall into one of two categories; either proprietary work of a consortium, or proprietary work of a individual/corporation.

Best Regards,
Joel

________________________________

[HP Circle No Bkgrbd]

Joel J. Fleck, II
Chief Architect
Head of Standards
HP Software CTO
Hewlett-Packard Company
200 Connell Parkway, Suite 500
Berkeley Heights, NJ 07922

+1 908.898.4091 Phone
+1 908.416.1706 Mobile
+1 908.898.4887 Fax
joel.fleck@hp.com

________________________________

Kenneth Dilbeck , Director, Technical Program , TM Forum 8/16/2011 5:16:33 PM
Folks,
Anyone familiar with SADL http://sadl.sourceforge.net/?
Perhaps if both groups can use the same tool base, Eclipse, the chances of keeping them together would increase. The JOSIF work is Eclipse based as well.
It would seem having tool support for something as rigorous as an ontology would be helpful. I have only taken a quick look and this may be more oriented towards reasoning systems but thought I would ask.
We do need to make sure there is a very open line of communication between the FAST meta-model team and any other ontology work that takes places to avoid duplication effort and development of conflicting answers.
Regards,
Ken

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd. 8/16/2011 4:47:53 PM

Hi Wedge,

Isnt modelling itself an abstraction activity, therefore 2 and 1 are same?

Sitaraman

William Greene , Technical SME; CTR DES-E , DOD 8/16/2011 4:46:08 PM
Classification: UNCLASSIFIED
Caveats: NONE

I’m a little shy of using “automated reasoning” as a goal. I’m personally satisfied with “automated services” as a more reachable goal. Perhaps, "automated, adaptive services” as a stretch goal.

I'm simply happy to start with defining the grammar, syntax, semantics. Also, maybe a few basic patterns that support first understanding Frameworx as a model and then basic operations on information contained in the model.

Eventually, (but not soon ;-) we might get to a point where we can use this to, for example, “frame the set of available interface solutions for domain X and iterate across these to determine an optimal decision tree for function L: federate with the similar decision tree coming from Domain Y.

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

David Milham , Chief Architect , TM Forum 8/16/2011 4:30:55 PM
One needed to be on the call to hear the full background and at the presentation at TAW Baltimore
The key point made by Steve was that this was an example of an enterprise repository solution some aspects of which need to be embedded in Frameworx, Without such a metamodel SP like BT assert that is it difficult to practical evolve systems interfaces business services etc. And it is not primarily a technical issue it is a management and governance need on Frameworx.

Dave
Dave Milham
Chief Architect
t +44 5601 855561 m+44 7802 902451
edmilham@tmforum.org

From: Exploring New Project Ideas [mailto:newprojects@community.tmforum.org] On Behalf Of Sitaraman Wills
Sent: 16 August 2011 21:14
To: Dave Milham
Subject: [Exploring New Project Ideas] TMF Ontology SIG [#169800]

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.

"David Milham" wrote:

Hi Dave,

Need a favor,

The following document seems to be a MetaModel proposal from Steve Orobeck, this looks like more of a solution document than a list of problems bothering a service provider as you seem to have intended. Am i looking at the right document and if so a pointer to where the problem statements are would help.

If you want to see what is bothering at least one service provider you should look up this contribution.
http://www.tmforum.org/Community/groups/frameworxarchitecture/contributions.aspx?id=artf2509
which was originally presented at TAW Baltimore in the cross Frameworx meetings

The following document gives me a permission denied error, would much aprpeicate if it can be sent as an attachment.
Thanks much.
Sitaraman

http://collab.tmforum.org/integration/viewcvs/viewcvs.cgi/trunk/?root=business_services&system=exsy1002

but we discovered that the individual meta-models don’t line up as we would wish.
Hence we started a top down exercise to complement the prior approach.

What we need now is hand and bodies to work through the patterns we need to use and identify and rectify the gaps in the existing component frameworks.

Dave
Dave Milham
Chief Architect
t +44 5601 855561m+44 7802 902451
edmilham@tmforum.org

From:Information Framework (SID) [mailto:sid@community.tmforum.org] On Behalf Of Sitaraman Wills
Sent: 16 August 2011 19:05
To: Dave Milham
Subject: [Information Framework (SID)] TMF Ontology SIG [#169800]

Sitaraman Wills, Senior Software Engineer , NMSWorks Software Pvt. Ltd.

At the end fo the day, we are TMF first and then Ontologists (to various levels of expertise, from bystanders, through believers, via budding through to been-there-done-thats ).

Perhaps we must have another line of attack from the ServiceProviders side. We could ask operators running their networks day in and day out...tell me in plain english or you native language what are the things that bother you in your job.

There must be complete freedom in responding, no tracking along a direction with a set of 10 questions and the like.

Then we huddle together and extract semantics out of them. Then, we can have that comfortable feeling that, at the end of the day, what we come up with is going to make an important set of people that drive an all pervading industry, smile, courtesy TMF.

Sitaraman

Information Framework (SID)
Topics:General

Shared In:4 Communities

Exploring New Project Ideas
Topics:General

Shared In:4 Communities

Paul Hanke , Sr Professional Staff , Johns Hopkins University Applied Physics Lab 8/16/2011 4:15:48 PM
.. I'm not saying that this is what everyone thinks of when they hear the word "ontology", but here is what a deliverable could look like:

An extensible, machine-readable, and foundational logical theory of service provider operations that supports automated reasoning and automated interoperability even across arbitrary domain-specific extensions.

That is, unlike a data model expressed in XML Schema where the best a machine can do with instance data is simply ingore any extensions it is unfamiliar with, an ontology expressed in OWL opens the possibility that a machine can formally manipulate instance data (in a way that mimics human understanding) to include extensions that it has never seen before ... perhaps this is what the FAST team is trying to address with meta-models, I couldn't say ... but even if it is, at this formative stage I see room for both thrust areas to poke and prod at the envelope of the possible.

William Greene , Technical SME; CTR DES-E , DOD 8/16/2011 4:15:42 PM
Classification: UNCLASSIFIED
Caveats: NONE

From Stephen, " Also, the JOSIF open source project has define and implemented a meta-model for interfaces."

Brilliant work!

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

David Milham , Chief Architect , TM Forum 8/16/2011 4:08:59 PM
Steve

Correct thanks for reminding us of the location of that information

Dave
Dave Milham
Chief Architect
t +44 5601 855561 m+44 7802 902451
edmilham@tmforum.org

From: Exploring New Project Ideas [mailto:newprojects@community.tmforum.org] On Behalf Of Stephen Fratini
Sent: 16 August 2011 20:52
To: Dave Milham
Subject: [Exploring New Project Ideas] TMF Ontology SIG [#169800]

Stephen Fratini , Senior Engineer , Telcordia Technologies

Dave and All,

Also, the JOSIF open source project has define and implemented a meta-model for interfaces.
See the JOSIF wiki pages at
http://sourceforge.net/apps/mediawiki/openoss/index.php?title=Interface_Meta-Model

Steve

---
Stephen Fratini
Telcordia Technologies, Chief Strategy Office
Chair of the TM Forum Integration Program
Home Office: +1 732 962 1500 (I tend to be at this number more often)
Office: +1 732 699 2226
Work Cell: +1 732 853 2207
Email: sfratini@telcordia.com

Exploring New Project Ideas
Topics:General

Shared In:4 Communities

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd. 8/16/2011 4:06:43 PM

"David Milham" wrote:

Hi Dave,

Need a favor,

The following document seems to be a MetaModel proposal from Steve Orobeck, this looks like more of a solution document than a list of problems bothering a service provider as you seem to have intended. Am i looking at the right document and if so a pointer to where the problem statements are would help.

If you want to see what is bothering at least one service provider you should look up this contribution.
http://www.tmforum.org/Community/groups/frameworxarchitecture/contributions.aspx?id=artf2509
which was originally presented at TAW Baltimore in the cross Frameworx meetings

The following document gives me a permission denied error, would much aprpeicate if it can be sent as an attachment.
Thanks much.
Sitaraman

http://collab.tmforum.org/integration/viewcvs/viewcvs.cgi/trunk/?root=business_services&system=exsy1002

but we discovered that the individual meta-models don’t line up as we would wish.
Hence we started a top down exercise to complement the prior approach.

What we need now is hand and bodies to work through the patterns we need to use and identify and rectify the gaps in the existing component frameworks.

Dave
Dave Milham
Chief Architect
t +44 5601 855561m+44 7802 902451
edmilham@tmforum.org

From:Information Framework (SID) [mailto:sid@community.tmforum.org] On Behalf Of Sitaraman Wills
Sent: 16 August 2011 19:05
To: Dave Milham
Subject: [Information Framework (SID)] TMF Ontology SIG [#169800]

Sitaraman Wills, Senior Software Engineer , NMSWorks Software Pvt. Ltd.

At the end fo the day, we are TMF first and then Ontologists (to various levels of expertise, from bystanders, through believers, via budding through to been-there-done-thats ).

Perhaps we must have another line of attack from the ServiceProviders side. We could ask operators running their networks day in and day out...tell me in plain english or you native language what are the things that bother you in your job.

There must be complete freedom in responding, no tracking along a direction with a set of 10 questions and the like.

Then we huddle together and extract semantics out of them. Then, we can have that comfortable feeling that, at the end of the day, what we come up with is going to make an important set of people that drive an all pervading industry, smile, courtesy TMF.

Sitaraman

Information Framework (SID)
Topics:General

Shared In:4 Communities

William Greene , Technical SME; CTR DES-E , DOD 8/16/2011 4:02:35 PM
Classification: UNCLASSIFIED
Caveats: NONE

I think we here and also the FAST team needs to distinguish between meta-model as:
- a model of a model
- an abstraction of a model.

At least in ontology, it should be an abstraction of a model.

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

Stephen Fratini , Senior Engineer , Telcordia Technologies 8/16/2011 3:40:13 PM
Dave and All,

Also, the JOSIF open source project has define and implemented a meta-model for interfaces.
See the JOSIF wiki pages at
http://sourceforge.net/apps/mediawiki/openoss/index.php?title=Interface_Meta-Model

Steve

---
Stephen Fratini
Telcordia Technologies, Chief Strategy Office
Chair of the TM Forum Integration Program
Home Office: +1 732 962 1500 (I tend to be at this number more often)
Office: +1 732 699 2226
Work Cell: +1 732 853 2207
Email: sfratini@telcordia.com

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd. 8/16/2011 3:34:26 PM

Hi Dave,

Thanks so much for those concrete pointers, I am already committed to the interface aspects of TMF, but at any open opportunitues i find, here is an intact body with appendages to boot which will do its best to chip in.

My first priority in this thread will now be to look thro the docs you have pointed out. Timeout from me for now.

Thanks again.

Sitaraman

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd. 8/16/2011 3:29:18 PM

Hi Jerome,

I have unilaterally decided to go against your call, in spite of being a relative novice, i am jumping in. What are your thoughts on the following, any pointers also would help.

A model is a reflection of something fundamental in the system under study. Those fundamental truths we try to capture in the model.

However whenwe talk of behavior and lifecycle there is an element of "free will" here, it becomes more difficult to assign something too fundamental to it and if you can it probably belongs as a characteristic. So (i do realize process models are widely accepted now but still if i ask myself do i truly believe in such a thing i am not able to confidently say yes i do,)process models are too dignfied a name, they are more of compilations fo similar process instances than a dignified models that abstractts a reality.

To take the line of thought, what Wedge was pointing out, what would be the general opinion on seeing

Information as modelled

States as Characteristic each of them forming/conforming to an integral part of that model.

And behaviour as a function action on a domain set of states and range same set of states

And life cycle as an instance of a compostie/complex behavior chain, a chain of functions, i guess Mathematicians term this as functional composition...

Mathematical Trees (Graphs including hierarchical bidirectional graphs), Sets, Functions and functional analysis seem to me to one set of candidate fundamental constructs we could look at...Wedge what would your thoughs be too on this?

Sitaraman

David Milham , Chief Architect , TM Forum 8/16/2011 3:24:56 PM
For information

The FAST team is preparing a second pass at Meta-models for Frameworx.

The TR structure is at
http://collab.tmforum.org/sf/docman/do/listDocuments/projects.integration_framework/docman.root.metamodels.tr1xx

And the first contribution from BT based on an operational enterprise repository was made earlier this week and discussed on today’s FAST calls.

If you want to see what is bothering at least one service provider you should look up this contribution.
http://www.tmforum.org/Community/groups/frameworxarchitecture/contributions.aspx?id=artf2509
which was originally presented at TAW Baltimore in the cross Frameworx meetings

The first attempt was to capture the meta-models in the existing component frameworks . This bottom up approach has been competed using ECORE metamodelling language an dis on subversion at:

http://collab.tmforum.org/integration/viewcvs/viewcvs.cgi/trunk/?root=business_services&system=exsy1002

but we discovered that the individual meta-models don’t line up as we would wish.
Hence we started a top down exercise to complement the prior approach.

What we need now is hand and bodies to work through the patterns we need to use and identify and rectify the gaps in the existing component frameworks.

Dave
Dave Milham
Chief Architect
t +44 5601 855561 m+44 7802 902451
edmilham@tmforum.org

From: Information Framework (SID) [mailto:sid@community.tmforum.org] On Behalf Of Sitaraman Wills
Sent: 16 August 2011 19:05
To: Dave Milham
Subject: [Information Framework (SID)] TMF Ontology SIG [#169800]

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.

At the end fo the day, we are TMF first and then Ontologists (to various levels of expertise, from bystanders, through believers, via budding through to been-there-done-thats ).

Perhaps we must have another line of attack from the ServiceProviders side. We could ask operators running their networks day in and day out...tell me in plain english or you native language what are the things that bother you in your job.

There must be complete freedom in responding, no tracking along a direction with a set of 10 questions and the like.

Then we huddle together and extract semantics out of them. Then, we can have that comfortable feeling that, at the end of the day, what we come up with is going to make an important set of people that drive an all pervading industry, smile, courtesy TMF.

Sitaraman

Information Framework (SID)
Topics:General

Shared In:4 Communities

David Milham , Chief Architect , TM Forum 8/16/2011 3:04:40 PM
Jerome
That’s my opinion too. I like to understand better exactly what a deliverable would look like

Dave
Dave Milham
Chief Architect
t +44 5601 855561 m+44 7802 902451
edmilham@tmforum.org

From: Information Framework (SID) [mailto:sid@community.tmforum.org]
Sent: 16 August 2011 18:10
To: Dave Milham
Subject: [Information Framework (SID)] TMF Ontology SIG [#169800]

Jerome Hannebelle , Director, Open Innovation, FT/IMG/Technocentre/BP , France Telecom

Hi Pierre,

This is definitely an interesting topic, which I believe relates to FAST team's intent to define the Frameworx's Metamodel.

Jerome

Information Framework (SID)
Topics:General

Shared In:4 Communities

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd. 8/16/2011 2:54:34 PM

Ok, I hear you Wedge, what in your opinion would be those foundational structures we need to focus on,. I am shooting off the cuff there admitedly, any pointers to formal references would help. Now that i have had the benefit of free thiking on this subject without any set notions, thanks to you all i am ready to look at what has been formalized...

Sitaraman

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd. 8/16/2011 2:51:45 PM

Hi Wedge,

Glad you find the thought interesting... before putting on it a rather binding-sounding term such as requirement, how about we will be open, no tilt either way and see if when the chips fall in place self-similarity shows up. How many ever times we say it, we need a formal mathematic proof to associate strong property of a requirement, (those "necessary and sufficients" and the likes these mathematicians never find sufficient to use) that takes us into the relam of mathematics of mathematics truly self similar.

But as a lay man, Knowledge is something you can only get more effectively out of existing knoweldge, that is why as one progresses up the grades, one learns more using the tools of the existing knowledge, (let us forget for the moment the bing bang of the first baby cry conveniently) self similarity seems very likely.

Coming to think of it, it appears we do a breadth first, and as you go down the knowledge tree the no of child increases and specialist spend a lot of timein the numerous number of fine grained children of their parent field and lifetimes are not enough to go to the next sibling of the parent.

As an aside for the educationists then, what woud happen if people were given the choice for part depth-first education?! What would depth first and breadth first mean in the network semantics mapping we would come up with? For RCA with a high level fault at the top levels of the tree, we would likely need a BFS, for SIA it is DFS - Duality showing up again!

Sitaraman

William Greene , Technical SME; CTR DES-E , DOD 8/16/2011 2:34:17 PM
Classification: UNCLASSIFIED
Caveats: NONE

Sitaraman, I think I get the gist of this. But I feel you are jumping ahead here. Running too far down the inference road. While I agree, an ontology which opens the door for your conclusions also would be my goal (graph tree, semantic processing, parent/child service structures...); we are not there yet. And must rigorously work out the foundation to get there.

But fair disclosure: I am not an expert in ontology.

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

Stephen Estes , Business Process Architect , Cisco Systems 8/16/2011 2:28:34 PM

Yes, I'd be interested.

Currently working with NIST on their cloud ref arch ontology.

-- Stephen

William Greene , Technical SME; CTR DES-E , DOD 8/16/2011 2:24:04 PM
Classification: UNCLASSIFIED
Caveats: NONE

Yea, Jerome, not so easy ;-)

These are characterizing statements; perhaps true; but neither providing definitions nor establishing relationships.

For example, Behavior and Lifecycle, must be defined in the semantics of the Ontology. The metamodel executes a ontological syntax, a grammar, in applying behavior and Lifecycle to specific targets. (first pass)

Sitaraman, I think, raises an interesting supposition: is the Frameworx metamodel recursive or self referential? Should that be a structural requirement?

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd. 8/16/2011 1:57:06 PM

At the end fo the day, we are TMF first and then Ontologists (to various levels of expertise, from bystanders, through believers, via budding through to been-there-done-thats ).

Perhaps we must have another line of attack from the ServiceProviders side. We could ask operators running their networks day in and day out...tell me in plain english or you native language what are the things that bother you in your job.

There must be complete freedom in responding, no tracking along a direction with a set of 10 questions and the like.

Then we huddle together and extract semantics out of them. Then, we can have that comfortable feeling that, at the end of the day, what we come up with is going to make an important set of people that drive an all pervading industry, smile, courtesy TMF.

Sitaraman

Jerome Hannebelle , Director, Open Innovation, FT/IMG/Technocentre/BP , France Telecom 8/16/2011 1:49:04 PM

Model experts are certainly more qualified than I, but I would think :

1) While an ontology focuses on giving a meaning(= a semantic) to "objects" (sometimes not as formally as a metamodel), a metamodel probably goes beyond in also looking at behavior, lifecycle, ..

2) The metamodel may define abstract, unifying concepts that may not show up in an ontology (because they are abstract).

Mr experts, please jump in !

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd. 8/16/2011 1:34:32 PM

I am particularly interested (from the most generatl interface harmonization perspective) in the following goals.

- Can we structure Semantics (we need not consider the general programming language semantics, a perhaps more difficult problem, but can confine to the most frequently occuring, to begin with, semantic patterns in telecommunication networks) Sub goals would be then

a) What are the frequently occuring "network semantics" both across and within Networks. We can and perphaps must free ourselves of all constraints and feel free to to be abstract here. Then

b) Can we represent them in COMMON structural pattern, concretely well defined with as little ambiguity as possible.

- Now that we have a COMMON pattern, for the most popularly occuring (to keep things simple at first) concretizations seen above

a)Represent different instances of the same concrete Semantics in the common pattern and lay them out.

b)In the abstract (free form) how do we characterize the "extent of difference" between the two.

c)How do we concretize the differnce expression (procedures, algorithms on the said common structure) that unambigously characterizes the diff. That is, from the diff, one must be able to get to the other from the other, other and converse.

d) If we do reach c) then hopefully the holy grail of merge (All of A in B, or All of B in A or B = A+B or B = B-A and other possibilities) will be a simpler extension.

I am hoping, well i would go the extent of saying it is almost a hunch, that the common strcture would have aspects of self-simlarity. Therefore i am hoping the above operations can be recursively applied to whatever turns out to be the "Atomic Operation" on the said structure and therefore be able to do complex "semantic merges", "Semantic Morphs" or Semantc Overlays and such operations out of the recursive operations composed out of the atomic operations.

If we reach here we have set ourselves up to casting these a)Structure and b)Unit )perations in telecommunicational semnatic processing (due credits to the Chemical Engineering clan),

Then if only we get lucky and characterize state (a poorer cousin of full semantics perhaps) with the same pattern, if state is a simpler form of semantics it does seem likely,

-Record State of network entities, of groups of network entities (same state compositional types as above, we can reuse those)

- Give said state(s) as input to Unit Operations and its constituents, and based on the input call suitable child operations.... We have context aware, semantic processing,

- Now have state change as part of the child operation and there you have feedback control, with semantics and syntax, a partial learning system, now we can appeal fo more higher levels of concretized math, Stability, Pole-Zero Analysis and other such intriguing things with which Control Theorists work their systems....?

And here it is... I would much appreciate peoples time in responding why that common structure cannot be the simple TREE of arbitray depth and in general a system represented by a forest of such trees? Concrete examples where it fails in achieving a given functionality would help much

Thanks.

Sitaraman

William Greene , Technical SME; CTR DES-E , DOD 8/16/2011 1:23:22 PM
Classification: UNCLASSIFIED
Caveats: NONE

Jerome, could you expound on " believe relates to FAST team's intent to define the Frameworx's Metamodel." My gut feeling is you are right (that is, I am not disputing this comment at all). I feel this is an important statement; but I need some help seeing exactly how the ontology and the metamodel are linked.

From: Wedge Greene

Craig Gallen , Consultant OpenNMS group , The OpenNMS Group, Inc. 8/16/2011 1:08:15 PM

Hi Pierre,

As discussed, I am very interested in this and would like to include the ten University labs I represent through the web Science trust. www.webscience.org

Cheers

Craig

Jerome Hannebelle , Director, Open Innovation, FT/IMG/Technocentre/BP , France Telecom 8/16/2011 1:06:36 PM

Hi Pierre,

This is definitely an interesting topic, which I believe relates to FAST team's intent to define the Frameworx's Metamodel.

Jerome

William Greene , Technical SME; CTR DES-E , DOD 8/16/2011 1:02:58 PM
Classification: UNCLASSIFIED
Caveats: NONE

I'm interested. However, I think after establishing a basic syntax and some inclusive semantic patterns (e.g., search, aggregation, collaboration, policy), fleshing out the various semantic definitions can be handled in the many traditional workgroups. What we eventually need is for each team to feed their semantics into the overall ontology just like we currently feed the SID and eTOM.

My specific interest today is on how an ontology will work with policy.

I do not see we need to get it perfectly right the first time; just we get it close enough to do much of the work... old 80/20 rule.

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd. 8/16/2011 12:46:06 PM

Hi Pierre,

Meself in it, initially however i request i be granted aceess to discussions and confine myself to observing and learning by participitating in those discussions. I do hope that would be ok.

Thanks.

Sitaraman

Alex Zhdankin , Modeling Architect , Cisco Systems 8/16/2011 12:42:30 PM

Count me in.

Alex.

John Wilmes , Chief Technical Architect, Communications Sector , Progress Software 8/16/2011 12:27:07 PM

I would like to participate especially with reference to SID and eTOM.

John

Paul Hanke , Sr Professional Staff , Johns Hopkins University Applied Physics Lab 8/16/2011 12:27:00 PM
.. responding with interest :-)

Bob Natale , Chief Engineer, Joint NetOps & Cyber Defense , MITRE 8/16/2011 12:16:44 PM
Count me in, Pierre.

Pierre Gauthier , CEO and Principal Consultant , Oss Wave 8/16/2011 11:50:18 AM

I would like to see if there is enough interest for creating an Ontology SIG within TMF.

I sugest that the initial focus be on establishing the foundation for a TMF Frameworx Ontology.

It is important that the set of proposed ontologies provides complete coverage of the Frameworx. This involves conceptual modeling of eTOM, SID, TAM but also complete coverage of the Business Services and TIP.

Please reply to that post if you are interested in participating to the TMF Ontology SIG.

End of Discussion History

Full thread available online here: http://www.tmforum.org/community/groups/exploring_new_project_ideas/forum/p/169800/192377.aspx?utm_campaign=community&utm_medium=exploring_new_project_ideas&utm_source=alert#192377
%20New%20Project%20Ideas&utm_source=alert>

This message was sent to joel.fleck@hp.com and follows the subscription preferences of your membership of the Exploring New Project Ideas TM Forum Online Community.

To unsubscribe immediately, please click here and click the 'Edit Subscriptions' button from the Email Subscription preferences section located on the right hand side of the page just underneath the menu bar. To cut down on email volume, you may also choose to receive your discussion notifications via a daily or weekly digest email as opposed to immediate separate emails.

TM Forum, 240 Headquarters Plaza, East Tower, 10th Floor, Morristown, NJ 07960-6628, USA. Call +1 973 944 5100. info@tmforum.org
[http://www.tmforum.org/csemails/emailtrack5006772.gif]
Stephen Estes , Business Process Architect , Cisco Systems 8/17/2011 12:32:58 PM
Hi, Alex --

I'm on the list and following the discussion, which has quite a wide range of viewpoints.

-- Stephen

On Aug 16, 2011, at 4:13 PM, Alex Zhdankin wrote:
David Milham , Chief Architect , TM Forum 8/17/2011 11:57:55 AM
First it is clear to me that some ontology related activities would be of value. My concern is the original proposed scope.
I am much more comfortable with the later more focussed proposals – particularly that form Joel Fleck

A few specific comments
1.       Metamodels What we are doing with metamodels is far beyond what can be done with XSD – for example metamodels can force different kinds of cardinality constraints and impose cross element rules using OCL.

That being said the metamodel works is focused on supporting SP enterprise architecture needs which include open interfaces and enterprise architecture transformation. Based on the TAW Baltimore discussion on cross frameworks, EA use of Frameworx requires a documented metamodel to facilitate adoption of Component Frameworx by SPs.

2.       SID metamodel/Interface Metamodel group can have as a deliverable to come up with XSD equivalent of Models and Interfaces for which BTs metamodel doc could be a base.
What is the advantage of doing this other than in the SID or Interfaces group?

Dave
Dave Milham
Chief Architect
t  +44 5601 855561 m+44 7802 902451
edmilham@tmforum.org

From: Exploring New Project Ideas [mailto:newprojects@community.tmforum.org] On Behalf Of Sitaraman Wills
Sent: 16 August 2011 22:54
To: Dave Milham
Subject: [Exploring New Project Ideas] TMF Ontology SIG [#169800]

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
Dave/Jerome,
Yes, I would go with any set of deliverable as long as it is useful to the industry and those are clearly defined.
Having said that, i do feel that in a forum called Ontology SIG, the participants who expressed interest in this group as long as such a group exists should be allowed to go beyond a meta model.
A meta model, to me is the equivalent of an XSD. I helps validation, it checks for the grammar for the language spoken is correct, the statement could very well be to borrow from a poem,
It would be perfectly valid (context awareness, i am reminded of John Strassners post) to say "would like a footlong with everything on it please" in a blockbuster store. So we are back to the killed thread i am afraid, context aware policies and behaviors. But even that belongs in a thread for PBNM or PDNM,
But for a thread that calls itself Ontology SIG should be allowed to explore beyond policies and contexts. even.
Therefore a
a)SID metamodel/Interface Metamodel group can have as a deliverable to come up wth XSD equivalent of Models and Interfaces.for which BTs metamodel doc could be a base. Additional deliverables, automated consistency checks, both static when a new model entities or interfaces are defined(similar to strong typing carrier over to a higher layer here) up to fine grained interface signature/parameter level dynamic checking. We could even extend this to automatic translation between instances of two diferent meta models. About here the name justifies a stop.
b) PBNM which will deal with how to express policies and behaviours an context awareness and determining what are the typical policies that are useful.
c) Onotology SIG, where we go beyond metamodels and context awareness is there a fundametnal structure to knowledge representation, are there fundamental operations on those which would formalize knowledge evolution (collective knowledge, learning). I like Pauls definition, though it could turn out to be ambitious with a narrower focus to our domain by trading off loss in generality, we would have a surer sight of the goal.
Additional deliverables would be,
Exhaustively list concrete Network Semantic Instances.
Map problem space in telecommunication network operations to NEtwork Semantics
Specify using knowledge evolution as above for
   Interface harmonization without loss of semantics
       Simple RCA/SIA
       Self-*
       Trending and Planning.
If the feeling is above is too much to shoot for, i will go with it, but the group should cease to be called Ontology SIG, Ontology and its potential applications to Telecommunications Network
go beyond metamodeling is my feeling.
Sitaraman


Exploring New Project Ideas
Topics:General

Shared In:4 Communities
David Milham , Chief Architect , TM Forum 8/17/2011 11:42:08 AM
Joel,
This is a very helpful input and the areas you cite seems to be reasonably scoped activities for an Ontology SIG. My concern with the original proposal was that whilst we do need to look at ontology related matters the proposal as stated was very broad, ambitious and will compete for limited SP resources from current date committed activities.

Do you think it would be possible to get a presentation of the IEEE paper from either yourself or Sven?

Dave
Dave Milham
Chief Architect
t  +44 5601 855561 m+44 7802 902451
edmilham@tmforum.org

From: Exploring New Project Ideas [mailto:newprojects@community.tmforum.org] On Behalf Of Joel Fleck II
Sent: 16 August 2011 22:49
To: Dave Milham
Subject: [Exploring New Project Ideas] TMF Ontology SIG [#169800]

Joel Fleck II , Chief Architect , Hewlett-Packard
Wedge et al:

One of the key directions in research and industry looking forward is the use of a model-driven approach for the construction of business vocabularies, rules, ontologies and technology-neutral representations of the models. In particular, recent research has demonstrated a very usable approach using the OMG’s Semantics for Business Vocabulary and Rules (SBVR) as the basis for transformations to OWL 2.0 (for ontologies and ontological based reasoning –read that semantics), and as the basis for N3 or N-tuple technology-neutral based representations for the policy language(s) of choice, to UML and OCL for the generation of code stubs, and finally to XPATH 2.0 for testing potential interfaces. For reference, I refer to a paper in IEEE Communications August, 2010 “Multidomain IT Architectures for Next-Generation Communications Service Providers”.

Just as an FYI, there has been significant effort on the formulation of an SBVR core for the specification of inter/intra domain relationships and the management of these relationships (and the transformations of that work into UML, OWL, and N3). The methodology has been successfully proven and shown to be very valuable in capturing the business perspective and building software architecture models from them. I would strongly advise investigation into these approaches.

However, at the current time, it would be premature to suggest that the actual models (and a couple of the transformation engines) would be contributed to the TMF as these component parts currently fall into one of two categories; either proprietary work of a consortium, or proprietary work of a individual/corporation.

Best Regards,
Joel

Joel J. Fleck, II
Chief Architect
Head of Standards
HP Software CTO
Hewlett-Packard Company
200 Connell Parkway, Suite 500
Berkeley Heights, NJ 07922

+1 908.898.4091 Phone
+1 908.416.1706 Mobile
+1 908.898.4887 Fax
joel.fleck@hp.com

Exploring New Project Ideas
Topics:General

Shared In:4 Communities
Raghavendra Tadepalli , Architect , Wipro Technologies 8/17/2011 11:01:44 AM
Count me in for this.
Br,
Raghu.

Alex Zhdankin , Modeling Architect , Cisco Systems 8/16/2011 6:37:51 PM
One of the topics the TM Forum Ontology SIG may be focusing on or at least should consider is related to the problem of the information federation, which is central to the cross-SDO work on the model and interface federation/harmonization that we've started.
Some time ago we (TM Forum) has filed through the OMG Telecom PSIG an RFI on Using Problem Specific Views for Information Federation essentially asking for the methodology that would enable creation of the truly connected federated information ecosystem. Apparently the issues described in the RFI resonated quite well with work that been performed by the OMG Architecture Ecosystem SIG (AESIG). The AESIG sent us the response to the RFI referring to the Semantic Information Modeling for Federation (SIMF) RFP, which they've prepared.
Both RFI and the response can be found here:
http://collab.tmforum.org/sf/docman/do/listDocuments/projects.integration_framework/docman.root.tooling_task_force_documents.working_documents_for_end_to_end.omg_rfi
Here is the link to the OMG Wiki where the actual SIMF RFP as well as related discussions can be found:
http://www.omgwiki.org/architecture-ecosystem/doku.php?id=semantic_information_modeling_for_federation_rfp
Regards,
Alex.
Alex Zhdankin , Modeling Architect , Cisco Systems 8/16/2011 6:23:39 PM
One of the areas where the Ontology SIG may focus on ot at least consider is related to the issue of the information federation, which is central to the cross-SDO work on the model and iterface federation/harmonization.
Some time ago we (TM Forum) filed and RFI with OMG Telecom PSIG asking for the methodology allowing to define the problem-specific views for information federation. This RFI has resonated quite well with the work that already has been started by the OMG Architecture Ecosystem SIG(AESIG), which has sent the Semantic Information Modeling for Federation (SIMF) RFP as a response to our RFI.
You can find both the original RFI and the response here:
http://collab.tmforum.org/sf/docman/do/listDocuments/projects.integration_framework/docman.root.tooling_task_force_documents.working_documents_for_end_to_end.omg_rfi
The text of the SIMF RFP and the related discussions can be found here:
http://www.omgwiki.org/architecture-ecosystem/doku.php?id=semantic_information_modeling_for_federation_rfp

Regards,
Alex.
David Ingersoll , VP Sales , Versant Corporation (Corporate D) 8/16/2011 6:12:35 PM
All,

Would it be useful for us, Versant, to provide a persistence mechanism for
the ontology. If so, I would be glad to assign a resource to facilitate
working with the group.

David Ingersoll | VP, Sales Americas & APAC
Direct: (650) 232-2430 | EMail: dingersoll@versant.com

(Embedded image moved to file: pic20640.gif) (Embedded image
moved to file: pic26052.gif)
255 Shoreline Drive, Suite 450, Redwood City, CA 94065 | main:
650.232.2400 | fax: 650.232.2401 | www.versant.com | www.db4o.com

CONFIDENTIALITY NOTICE: This e-mail message, including any attachments, is
for the sole use of the intended recipient(s) and may contain confidential
or proprietary information. Any unauthorized review, use, disclosure or
distribution is prohibited. If you are not the intended recipient,
immediately contact the sender by reply e-mail and destroy all copies of
the original message

From: "Joel Fleck II"

To: dingersoll@versant.com

Date: 08/16/2011 02:51 PM

Subject: [Information Framework (SID)] TMF Ontology SIG [#169800]

Sent by: "Information Framework (SID)"

Joel Fleck II , Chief Architect , Hewlett-Packard

Wedge et al:

One of the key directions in research and industry looking forward is the use of a model-driven approach for the construction of business
vocabularies, rules, ontologies and technology-neutral representations of the models. In particular, recent research has demonstrated a very usable
approach using the OMG’s Semantics for Business Vocabulary and Rules (SBVR) as the basis for transformations to OWL 2.0 (for ontologies and
ontological based reasoning –read that semantics), and as the basis for N3 or N-tuple technology-neutral based representations for the policy
language(s) of choice, to UML and OCL for the generation of code stubs, and finally to XPATH 2.0 for testing potential interfaces. For reference, I
refer to a paper in IEEE Communications August, 2010 “Multidomain IT Architectures for Next-Generation Communications Service Providers”.

Just as an FYI, there has been significant effort on the formulation of an SBVR core for the specification of inter/intra domain relationships and
the management of these relationships (and the transformations of that work into UML, OWL, and N3). The methodology has been successfully proven and
shown to be very valuable in capturing the business perspective and building software architecture models from them. I would strongly advise
investigation into these approaches.

However, at the current time, it would be premature to suggest that the actual models (and a couple of the transformation engines) would be
contributed to the TMF as these component parts currently fall into one of two categories; either proprietary work of a consortium, or proprietary
work of a individual/corporation.

Best Regards,

Joel

HP Circle No Bkgrbd
Joel J. Fleck, II
Chief Architect

Head of Standards

HP Software CTO
Hewlett-Packard Company
200 Connell Parkway, Suite 500

Berkeley Heights, NJ 07922

+1 908.898.4091 Phone
+1 908.416.1706 Mobile
+1 908.898.4887 Fax
joel.fleck@hp.com

Information Framework (SID)
Topics:General
Shared In:4 Communities
Shared In:Business Process Framework (eTOM) : General
TM Forum Integration Program (TIP) : General
Frameworx Architecture Steering Team (FAST) : General
Exploring New Project Ideas : General

View thread online

Please note, this discussion thread is shared across communities. Any replies that you make to this discussion thread, will be posted to all of the
communities and topics listed above.

Discussion History

Stephen Estes , Business Process Architect , Cisco Systems
8/16/2011 5:48:11 PM

OWL 2.0 is a current standard, and its subset, OWL-DL is both very expressive and logically tractable. For this, the editing tool, Protege 4.1x, is
open source, multi-platform (Java) and robustly maintained.

http://protege.stanford.edu/

-- Stephen

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 5:46:19 PM

And yes, as Ken says one group can feed into the other so the whole becomes greater than the parts with team focused on clearly defined agendas and
work in the same environment so that there not avodable software-administrrative hassles that hurt progress. In that sense a Metamodel is indeed an
important first step, we need to have our grammar straight before we use it to consistently communicate semantics without chaos due to
misinterpretation.

Sitaraman

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 5:42:11 PM

Dave/Jerome,

Yes, I would go with any set of deliverable as long as it is useful to the industry and those are clearly defined.

Having said that, i do feel that in a forum called Ontology SIG, the participants who expressed interest in this group as long as such a group
exists should be allowed to go beyond a meta model.

A meta model, to me is the equivalent of an XSD. I helps validation, it checks for the grammar for the language spoken is correct, the statement
could very well be to borrow from a poem,

It would be perfectly valid (context awareness, i am reminded of John Strassners post) to say "would like a footlong with everything on it
please" in a blockbuster store. So we are back to the killed thread i am afraid, context aware policies and behaviors. But even that belongs in a
thread for PBNM or PDNM,

But for a thread that calls itself Ontology SIG should be allowed to explore beyond policies and contexts. even.

Therefore a

a)SID metamodel/Interface Metamodel group can have as a deliverable to come up wth XSD equivalent of Models and Interfaces.for which BTs metamodel
doc could be a base. Additional deliverables, automated consistency checks, both static when a new model entities or interfaces are defined(similar
to strong typing carrier over to a higher layer here) up to fine grained interface signature/parameter level dynamic checking. We could even extend
this to automatic translation between instances of two diferent meta models. About here the name justifies a stop.

b) PBNM which will deal with how to express policies and behaviours an context awareness and determining what are the typical policies that are
useful.

c) Onotology SIG, where we go beyond metamodels and context awareness is there a fundametnal structure to knowledge representation, are there
fundamental operations on those which would formalize knowledge evolution (collective knowledge, learning). I like Pauls definition, though it could
turn out to be ambitious with a narrower focus to our domain by trading off loss in generality, we would have a surer sight of the goal.

Additional deliverables would be,

Exhaustively list concrete Network Semantic Instances.

Map problem space in telecommunication network operations to NEtwork Semantics

Specify using knowledge evolution as above for

Interface harmonization without loss of semantics

Simple RCA/SIA

Self-*

Trending and Planning.

If the feeling is above is too much to shoot for, i will go with it, but the group should cease to be called Ontology SIG, Ontology and its
potential applications to Telecommunications Network
go beyond metamodeling is my feeling.

Sitaraman

Kenneth Dilbeck , Director, Technical Program , TM Forum
8/16/2011 5:16:33 PM

Folks,

Anyone familiar with SADL http://sadl.sourceforge.net/?

Perhaps if both groups can use the same tool base, Eclipse, the chances of keeping them together would increase. The JOSIF work is Eclipse based as
well.

It would seem having tool support for something as rigorous as an ontology would be helpful. I have only taken a quick look and this may be more
oriented towards reasoning systems but thought I would ask.

We do need to make sure there is a very open line of communication between the FAST meta-model team and any other ontology work that takes places to
avoid duplication effort and development of conflicting answers.

Regards,

Ken

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 4:47:53 PM

Hi Wedge,

Isnt modelling itself an abstraction activity, therefore 2 and 1 are same?

Sitaraman

William Greene , Technical SME; CTR DES-E , DOD
8/16/2011 4:46:08 PM

Classification: UNCLASSIFIED
Caveats: NONE

I’m a little shy of using “automated reasoning” as a goal. I’m personally satisfied with “automated services” as a more reachable goal. Perhaps,
"automated, adaptive services” as a stretch goal.

I'm simply happy to start with defining the grammar, syntax, semantics. Also, maybe a few basic patterns that support first understanding Frameworx
as a model and then basic operations on information contained in the model.

Eventually, (but not soon ;-) we might get to a point where we can use this to, for example, “frame the set of available interface solutions for
domain X and iterate across these to determine an optimal decision tree for function L: federate with the similar decision tree coming from Domain
Y.

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

David Milham , Chief Architect , TM Forum
8/16/2011 4:30:55 PM

One needed to be on the call to hear the full background and at the presentation at TAW Baltimore

The key point made by Steve was that this was an example of an enterprise repository solution some aspects of which need to be embedded in
Frameworx, Without such a metamodel SP like BT assert that is it difficult to practical evolve systems interfaces business services etc. And it is
not primarily a technical issue it is a management and governance need on Frameworx.

Dave

Dave Milham

Chief Architect

t +44 5601 855561 m+44 7802 902451
edmilham@tmforum.org

From: Exploring New Project Ideas [mailto:newprojects@community.tmforum.org] On Behalf Of Sitaraman Wills
Sent: 16 August 2011 21:14
To: Dave Milham
Subject: [Exploring New Project Ideas] TMF Ontology SIG [#169800]

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.

"David Milham" wrote:

Hi Dave,

Need a favor,

The following document seems to be a MetaModel proposal from Steve Orobeck, this looks like more of a solution document than a list of problems
bothering a service provider as you seem to have intended. Am i looking at the right document and if so a pointer to where the problem
statements are would help.

If you want to see what is bothering at least one service provider you should look up this contribution.

http://www.tmforum.org/Community/groups/frameworxarchitecture/contributions.aspx?id=artf2509

which was originally presented at TAW Baltimore in the cross Frameworx meetings

The following document gives me a permission denied error, would much aprpeicate if it can be sent as an attachment.

Thanks much.

Sitaraman

http://collab.tmforum.org/integration/viewcvs/viewcvs.cgi/trunk/?root=business_services&system=exsy1002

but we discovered that the individual meta-models don’t line up as we would wish.

Hence we started a top down exercise to complement the prior approach.

What we need now is hand and bodies to work through the patterns we need to use and identify and rectify the gaps in the existing component
frameworks.

Dave

Dave Milham

Chief Architect

t +44 5601 855561m+44 7802 902451
edmilham@tmforum.org

From:Information Framework (SID) [mailto:sid@community.tmforum.org] On Behalf Of Sitaraman Wills
Sent: 16 August 2011 19:05
To: Dave Milham
Subject: [Information Framework (SID)] TMF Ontology SIG [#169800]

Sitaraman Wills, Senior Software Engineer , NMSWorks Software Pvt. Ltd.

At the end fo the day, we are TMF first and then Ontologists (to various levels of expertise, from bystanders, through believers, via budding
through to been-there-done-thats ).

Perhaps we must have another line of attack from the ServiceProviders side. We could ask operators running their networks day in and day out...tell
me in plain english or you native language what are the things that bother you in your job.

There must be complete freedom in responding, no tracking along a direction with a set of 10 questions and the like.

Then we huddle together and extract semantics out of them. Then, we can have that comfortable feeling that, at the end of the day, what we come up
with is going to make an important set of people that drive an all pervading industry, smile, courtesy TMF.

Sitaraman
Information Framework (SID)
Topics:General

Shared In:4 Communities

Exploring New Project Ideas
Topics:General

Shared In:4 Communities

Paul Hanke , Sr Professional Staff , Johns Hopkins University Applied Physics Lab
8/16/2011 4:15:48 PM

.. I'm not saying that this is what everyone thinks of when they hear the word "ontology", but here is what a deliverable could look like:

An extensible, machine-readable, and foundational logical theory of service provider operations that supports automated reasoning and automated
interoperability even across arbitrary domain-specific extensions.

That is, unlike a data model expressed in XML Schema where the best a machine can do with instance data is simply ingore any extensions it is
unfamiliar with, an ontology expressed in OWL opens the possibility that a machine can formally manipulate instance data (in a way that mimics human
understanding) to include extensions that it has never seen before .. perhaps this is what the FAST team is trying to address with meta-models, I
couldn't say ... but even if it is, at this formative stage I see room for both thrust areas to poke and prod at the envelope of the possible.

William Greene , Technical SME; CTR DES-E , DOD
8/16/2011 4:15:42 PM

Classification: UNCLASSIFIED
Caveats: NONE

From Stephen, " Also, the JOSIF open source project has define and implemented a meta-model for interfaces."

Brilliant work!

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

David Milham , Chief Architect , TM Forum
8/16/2011 4:08:59 PM

Steve

Correct thanks for reminding us of the location of that information

Dave

Dave Milham

Chief Architect

t +44 5601 855561 m+44 7802 902451
edmilham@tmforum.org

From: Exploring New Project Ideas [mailto:newprojects@community.tmforum.org] On Behalf Of Stephen Fratini
Sent: 16 August 2011 20:52
To: Dave Milham
Subject: [Exploring New Project Ideas] TMF Ontology SIG [#169800]

Stephen Fratini , Senior Engineer , Telcordia Technologies

Dave and All,

Also, the JOSIF open source project has define and implemented a meta-model for interfaces.

See the JOSIF wiki pages at

http://sourceforge.net/apps/mediawiki/openoss/index.php?title=Interface_Meta-Model

Steve

---

Stephen Fratini

Telcordia Technologies, Chief Strategy Office

Chair of the TM Forum Integration Program

Home Office: +1 732 962 1500 (I tend to be at this number more often)

Office: +1 732 699 2226

Work Cell: +1 732 853 2207

Email: sfratini@telcordia.com

Exploring New Project Ideas
Topics:General

Shared In:4 Communities

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 4:06:43 PM

"David Milham" wrote:

Hi Dave,

Need a favor,

The following document seems to be a MetaModel proposal from Steve Orobeck, this looks like more of a solution document than a list of problems
bothering a service provider as you seem to have intended. Am i looking at the right document and if so a pointer to where the problem
statements are would help.

If you want to see what is bothering at least one service provider you should look up this contribution.

http://www.tmforum.org/Community/groups/frameworxarchitecture/contributions.aspx?id=artf2509

which was originally presented at TAW Baltimore in the cross Frameworx meetings

The following document gives me a permission denied error, would much aprpeicate if it can be sent as an attachment.

Thanks much.

Sitaraman

http://collab.tmforum.org/integration/viewcvs/viewcvs.cgi/trunk/?root=business_services&system=exsy1002

but we discovered that the individual meta-models don’t line up as we would wish.

Hence we started a top down exercise to complement the prior approach.

What we need now is hand and bodies to work through the patterns we need to use and identify and rectify the gaps in the existing component
frameworks.

Dave

Dave Milham

Chief Architect

t +44 5601 855561m+44 7802 902451
edmilham@tmforum.org

From:Information Framework (SID) [mailto:sid@community.tmforum.org] On Behalf Of Sitaraman Wills
Sent: 16 August 2011 19:05
To: Dave Milham
Subject: [Information Framework (SID)] TMF Ontology SIG [#169800]

Sitaraman Wills, Senior Software Engineer , NMSWorks Software Pvt. Ltd.

At the end fo the day, we are TMF first and then Ontologists (to various levels of expertise, from bystanders, through believers, via budding
through to been-there-done-thats ).

Perhaps we must have another line of attack from the ServiceProviders side. We could ask operators running their networks day in and day out...tell
me in plain english or you native language what are the things that bother you in your job.

There must be complete freedom in responding, no tracking along a direction with a set of 10 questions and the like.

Then we huddle together and extract semantics out of them. Then, we can have that comfortable feeling that, at the end of the day, what we come up
with is going to make an important set of people that drive an all pervading industry, smile, courtesy TMF.

Sitaraman
Information Framework (SID)
Topics:General

Shared In:4 Communities

William Greene , Technical SME; CTR DES-E , DOD
8/16/2011 4:02:35 PM

Classification: UNCLASSIFIED
Caveats: NONE

I think we here and also the FAST team needs to distinguish between meta-model as:
- a model of a model
- an abstraction of a model.

At least in ontology, it should be an abstraction of a model.

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

Stephen Fratini , Senior Engineer , Telcordia Technologies
8/16/2011 3:40:13 PM

Dave and All,

Also, the JOSIF open source project has define and implemented a meta-model for interfaces.
See the JOSIF wiki pages at
http://sourceforge.net/apps/mediawiki/openoss/index.php?title=Interface_Meta-Model

Steve

---
Stephen Fratini
Telcordia Technologies, Chief Strategy Office
Chair of the TM Forum Integration Program
Home Office: +1 732 962 1500 (I tend to be at this number more often)
Office: +1 732 699 2226
Work Cell: +1 732 853 2207
Email: sfratini@telcordia.com

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 3:34:26 PM

Hi Dave,

Thanks so much for those concrete pointers, I am already committed to the interface aspects of TMF, but at any open opportunitues i find, here is
an intact body with appendages to boot which will do its best to chip in.

My first priority in this thread will now be to look thro the docs you have pointed out. Timeout from me for now.

Thanks again.

Sitaraman

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 3:29:18 PM

Hi Jerome,

I have unilaterally decided to go against your call, in spite of being a relative novice, i am jumping in. What are your thoughts on the following,
any pointers also would help.

A model is a reflection of something fundamental in the system under study. Those fundamental truths we try to capture in the model.

However whenwe talk of behavior and lifecycle there is an element of "free will" here, it becomes more difficult to assign something too
fundamental to it and if you can it probably belongs as a characteristic. So (i do realize process models are widely accepted now but still if i ask
myself do i truly believe in such a thing i am not able to confidently say yes i do,)process models are too dignfied a name, they are more of
compilations fo similar process instances than a dignified models that abstractts a reality.

To take the line of thought, what Wedge was pointing out, what would be the general opinion on seeing

Information as modelled

States as Characteristic each of them forming/conforming to an integral part of that model.

And behaviour as a function action on a domain set of states and range same set of states

And life cycle as an instance of a compostie/complex behavior chain, a chain of functions, i guess Mathematicians term this as functional
composition...

Mathematical Trees (Graphs including hierarchical bidirectional graphs), Sets, Functions and functional analysis seem to me to one set of
candidate fundamental constructs we could look at...Wedge what would your thoughs be too on this?

Sitaraman

David Milham , Chief Architect , TM Forum
8/16/2011 3:24:56 PM

For information

The FAST team is preparing a second pass at Meta-models for Frameworx.

The TR structure is at

http://collab.tmforum.org/sf/docman/do/listDocuments/projects.integration_framework/docman.root.metamodels.tr1xx

And the first contribution from BT based on an operational enterprise repository was made earlier this week and discussed on today’s FAST calls.

If you want to see what is bothering at least one service provider you should look up this contribution.

http://www.tmforum.org/Community/groups/frameworxarchitecture/contributions.aspx?id=artf2509

which was originally presented at TAW Baltimore in the cross Frameworx meetings

The first attempt was to capture the meta-models in the existing component frameworks . This bottom up approach has been competed using ECORE
metamodelling language an dis on subversion at:

http://collab.tmforum.org/integration/viewcvs/viewcvs.cgi/trunk/?root=business_services&system=exsy1002

but we discovered that the individual meta-models don’t line up as we would wish.

Hence we started a top down exercise to complement the prior approach.

What we need now is hand and bodies to work through the patterns we need to use and identify and rectify the gaps in the existing component
frameworks.

Dave

Dave Milham

Chief Architect

t +44 5601 855561 m+44 7802 902451
edmilham@tmforum.org

From: Information Framework (SID) [mailto:sid@community.tmforum.org] On Behalf Of Sitaraman Wills
Sent: 16 August 2011 19:05
To: Dave Milham
Subject: [Information Framework (SID)] TMF Ontology SIG [#169800]

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.

At the end fo the day, we are TMF first and then Ontologists (to various levels of expertise, from bystanders, through believers, via budding
through to been-there-done-thats ).

Perhaps we must have another line of attack from the ServiceProviders side. We could ask operators running their networks day in and day out...tell
me in plain english or you native language what are the things that bother you in your job.

There must be complete freedom in responding, no tracking along a direction with a set of 10 questions and the like.

Then we huddle together and extract semantics out of them. Then, we can have that comfortable feeling that, at the end of the day, what we come up
with is going to make an important set of people that drive an all pervading industry, smile, courtesy TMF.

Sitaraman
Information Framework (SID)
Topics:General

Shared In:4 Communities

David Milham , Chief Architect , TM Forum
8/16/2011 3:04:40 PM

Jerome

That’s my opinion too. I like to understand better exactly what a deliverable would look like

Dave

Dave Milham

Chief Architect

t +44 5601 855561 m+44 7802 902451
edmilham@tmforum.org

From: Information Framework (SID) [mailto:sid@community.tmforum.org]
Sent: 16 August 2011 18:10
To: Dave Milham
Subject: [Information Framework (SID)] TMF Ontology SIG [#169800]

Jerome Hannebelle , Director, Open Innovation, FT/IMG/Technocentre/BP , France Telecom

Hi Pierre,

This is definitely an interesting topic, which I believe relates to FAST team's intent to define the Frameworx's Metamodel.

Jerome
Information Framework (SID)
Topics:General

Shared In:4 Communities

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 2:54:34 PM

Ok, I hear you Wedge, what in your opinion would be those foundational structures we need to focus on,. I am shooting off the cuff there admitedly,
any pointers to formal references would help. Now that i have had the benefit of free thiking on this subject without any set notions, thanks to you
all i am ready to look at what has been formalized...

Sitaraman

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 2:51:45 PM

Hi Wedge,

Glad you find the thought interesting... before putting on it a rather binding-sounding term such as requirement, how about we will be open, no
tilt either way and see if when the chips fall in place self-similarity shows up. How many ever times we say it, we need a formal mathematic proof
to associate strong property of a requirement, (those "necessary and sufficients" and the likes these mathematicians never find sufficient to use)
that takes us into the relam of mathematics of mathematics truly self similar.

But as a lay man, Knowledge is something you can only get more effectively out of existing knoweldge, that is why as one progresses up the grades,
one learns more using the tools of the existing knowledge, (let us forget for the moment the bing bang of the first baby cry conveniently) self
similarity seems very likely.

Coming to think of it, it appears we do a breadth first, and as you go down the knowledge tree the no of child increases and specialist spend a lot
of timein the numerous number of fine grained children of their parent field and lifetimes are not enough to go to the next sibling of the parent.

As an aside for the educationists then, what woud happen if people were given the choice for part depth-first education?! What would depth first
and breadth first mean in the network semantics mapping we would come up with? For RCA with a high level fault at the top levels of the tree, we
would likely need a BFS, for SIA it is DFS - Duality showing up again!

Sitaraman

William Greene , Technical SME; CTR DES-E , DOD
8/16/2011 2:34:17 PM

Classification: UNCLASSIFIED
Caveats: NONE

Sitaraman, I think I get the gist of this. But I feel you are jumping ahead here. Running too far down the inference road. While I agree, an
ontology which opens the door for your conclusions also would be my goal (graph tree, semantic processing, parent/child service structures...); we
are not there yet. And must rigorously work out the foundation to get there.

But fair disclosure: I am not an expert in ontology.

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

Stephen Estes , Business Process Architect , Cisco Systems
8/16/2011 2:28:34 PM

Yes, I'd be interested.

Currently working with NIST on their cloud ref arch ontology.

-- Stephen

William Greene , Technical SME; CTR DES-E , DOD
8/16/2011 2:24:04 PM

Classification: UNCLASSIFIED
Caveats: NONE

Yea, Jerome, not so easy ;-)

These are characterizing statements; perhaps true; but neither providing definitions nor establishing relationships.

For example, Behavior and Lifecycle, must be defined in the semantics of the Ontology. The metamodel executes a ontological syntax, a grammar, in
applying behavior and Lifecycle to specific targets. (first pass)

Sitaraman, I think, raises an interesting supposition: is the Frameworx metamodel recursive or self referential? Should that be a structural
requirement?

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 1:57:06 PM

At the end fo the day, we are TMF first and then Ontologists (to various levels of expertise, from bystanders, through believers, via budding
through to been-there-done-thats ).

Perhaps we must have another line of attack from the ServiceProviders side. We could ask operators running their networks day in and day out...tell
me in plain english or you native language what are the things that bother you in your job.

There must be complete freedom in responding, no tracking along a direction with a set of 10 questions and the like.

Then we huddle together and extract semantics out of them. Then, we can have that comfortable feeling that, at the end of the day, what we come up
with is going to make an important set of people that drive an all pervading industry, smile, courtesy TMF.

Sitaraman

Jerome Hannebelle , Director, Open Innovation, FT/IMG/Technocentre/BP , France Telecom
8/16/2011 1:49:04 PM

Model experts are certainly more qualified than I, but I would think :

1) While an ontology focuses on giving a meaning(= a semantic) to "objects" (sometimes not as formally as a metamodel), a metamodel probably goes
beyond in also looking at behavior, lifecycle, ...

2) The metamodel may define abstract, unifying concepts that may not show up in an ontology (because they are abstract).

Mr experts, please jump in !

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 1:34:32 PM

I am particularly interested (from the most generatl interface harmonization perspective) in the following goals.

- Can we structure Semantics (we need not consider the general programming language semantics, a perhaps more difficult problem, but can confine
to the most frequently occuring, to begin with, semantic patterns in telecommunication networks) Sub goals would be then

a) What are the frequently occuring "network semantics" both across and within Networks. We can and perphaps must free ourselves of all
constraints and feel free to to be abstract here. Then

b) Can we represent them in COMMON structural pattern, concretely well defined with as little ambiguity as possible.

- Now that we have a COMMON pattern, for the most popularly occuring (to keep things simple at first) concretizations seen above

a)Represent different instances of the same concrete Semantics in the common pattern and lay them out.

b)In the abstract (free form) how do we characterize the "extent of difference" between the two.

c)How do we concretize the differnce expression (procedures, algorithms on the said common structure) that unambigously characterizes the
diff. That is, from the diff, one must be able to get to the other from the other, other and converse.

d) If we do reach c) then hopefully the holy grail of merge (All of A in B, or All of B in A or B = A+B or B = B-A and other possibilities)
will be a simpler extension.

I am hoping, well i would go the extent of saying it is almost a hunch, that the common strcture would have aspects of self-simlarity.
Therefore i am hoping the above operations can be recursively applied to whatever turns out to be the "Atomic Operation" on the said structure and
therefore be able to do complex "semantic merges", "Semantic Morphs" or Semantc Overlays and such operations out of the recursive operations
composed out of the atomic operations.

If we reach here we have set ourselves up to casting these a)Structure and b)Unit )perations in telecommunicational semnatic processing (due
credits to the Chemical Engineering clan),

Then if only we get lucky and characterize state (a poorer cousin of full semantics perhaps) with the same pattern, if state is a simpler form
of semantics it does seem likely,

-Record State of network entities, of groups of network entities (same state compositional types as above, we can reuse those)

- Give said state(s) as input to Unit Operations and its constituents, and based on the input call suitable child operations.... We have context
aware, semantic processing,

- Now have state change as part of the child operation and there you have feedback control, with semantics and syntax, a partial learning system,
now we can appeal fo more higher levels of concretized math, Stability, Pole-Zero Analysis and other such intriguing things with which Control
Theorists work their systems....?

And here it is... I would much appreciate peoples time in responding why that common structure cannot be the simple TREE of arbitray depth and in
general a system represented by a forest of such trees? Concrete examples where it fails in achieving a given functionality would help much

Thanks.

Sitaraman

William Greene , Technical SME; CTR DES-E , DOD
8/16/2011 1:23:22 PM

Classification: UNCLASSIFIED
Caveats: NONE

Jerome, could you expound on " believe relates to FAST team's intent to define the Frameworx's Metamodel." My gut feeling is you are right (that is,
I am not disputing this comment at all). I feel this is an important statement; but I need some help seeing exactly how the ontology and the
metamodel are linked.

From: Wedge Greene

Craig Gallen , Consultant OpenNMS group , The OpenNMS Group, Inc.
8/16/2011 1:08:15 PM

Hi Pierre,

As discussed, I am very interested in this and would like to include the ten University labs I represent through the web Science trust.
www.webscience.org

Cheers

Craig

Jerome Hannebelle , Director, Open Innovation, FT/IMG/Technocentre/BP , France Telecom
8/16/2011 1:06:36 PM

Hi Pierre,

This is definitely an interesting topic, which I believe relates to FAST team's intent to define the Frameworx's Metamodel.

Jerome

William Greene , Technical SME; CTR DES-E , DOD
8/16/2011 1:02:58 PM

Classification: UNCLASSIFIED
Caveats: NONE

I'm interested. However, I think after establishing a basic syntax and some inclusive semantic patterns (e.g., search, aggregation, collaboration,
policy), fleshing out the various semantic definitions can be handled in the many traditional workgroups. What we eventually need is for each team
to feed their semantics into the overall ontology just like we currently feed the SID and eTOM.

My specific interest today is on how an ontology will work with policy.

I do not see we need to get it perfectly right the first time; just we get it close enough to do much of the work... old 80/20 rule.

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
8/16/2011 12:46:06 PM

Hi Pierre,

Meself in it, initially however i request i be granted aceess to discussions and confine myself to observing and learning by participitating in
those discussions. I do hope that would be ok.

Thanks.

Sitaraman

Alex Zhdankin , Modeling Architect , Cisco Systems
8/16/2011 12:42:30 PM

Count me in.

Alex.

John Wilmes , Chief Technical Architect, Communications Sector , Progress Software
8/16/2011 12:27:07 PM

I would like to participate especially with reference to SID and eTOM.

John

Paul Hanke , Sr Professional Staff , Johns Hopkins University Applied Physics Lab
8/16/2011 12:27:00 PM

.. responding with interest :-)

Bob Natale , Chief Engineer, Joint NetOps & Cyber Defense , MITRE
8/16/2011 12:16:44 PM

Count me in, Pierre.

Pierre Gauthier , CEO and Principal Consultant , Oss Wave
8/16/2011 11:50:18 AM

I would like to see if there is enough interest for creating an Ontology SIG within TMF.

I sugest that the initial focus be on establishing the foundation for a TMF Frameworx Ontology.

It is important that the set of proposed ontologies provides complete coverage of the Frameworx. This involves conceptual modeling of eTOM, SID, TAM
but also complete coverage of the Business Services and TIP.

Please reply to that post if you are interested in participating to the TMF Ontology SIG.

End of Discussion History

Full thread available online here:
http://www.tmforum.org/community/groups/information_framework_sid/forum/p/169800/192332.aspx?utm_campaign=community&utm_medium=information_framework
_sid&utm_source=alert#192332

----------------------------------------------------------------------------------------------------------------------------------------------------
This message was sent to dingersoll@versant.com and follows the subscription preferences of your membership of the Information Framework (SID) TM
Forum Online Community.

To unsubscribe immediately, please click here and click the 'Edit Subscriptions' button from the Email Subscription preferences section located on
the right hand side of the page just underneath the menu bar. To cut down on email volume, you may also choose to receive your discussion
notifications via a daily or weekly digest email as opposed to immediate separate emails.

TM Forum, 240 Headquarters Plaza, East Tower, 10th Floor, Morristown, NJ 07960-6628, USA. Call +1 973 944 5100. info@tmforum.org
Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd. 8/16/2011 5:54:44 PM
"However, at the current time, it would be premature to suggest that the actual models (and a couple of the transformation engines) would be contributed to the TMF as these component parts currently fall into one of two categories; either proprietary work of a consortium, or proprietary work of a individual/corporation."

Ah!
Would be most interested in TMFs opinion on this as a Forum/Consortium on this point. I agree with Joel, we say no this is not going to happen for various reaons and do other less ambitious but yet useful things as a group, or we say yes we can make this happen and put in the time and effort to come up with work. (New IPR though i understand very little of that document ) it sounds is intended to encourage the latter sort of thing?
Sitaraman
Stephen Estes , Business Process Architect , Cisco Systems 8/16/2011 5:48:11 PM
OWL 2.0 is a current standard, and its subset, OWL-DL is both very expressive and logically tractable. For this, the editing tool, Protege 4.1x, is open source, multi-platform (Java) and robustly maintained.
http://protege.stanford.edu/
-- Stephen

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd. 8/16/2011 5:46:19 PM
And yes, as Ken says one group can feed into the other so the whole becomes greater than the parts with team focused on clearly defined agendas and work in the same environment so that there not avodable software-administrrative hassles that hurt progress. In that sense a Metamodel is indeed an important first step, we need to have our grammar straight before we use it to consistently communicate semantics without chaos due to misinterpretation.
Sitaraman
Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd. 8/16/2011 5:42:11 PM
Dave/Jerome,
Yes, I would go with any set of deliverable as long as it is useful to the industry and those are clearly defined.
Having said that, i do feel that in a forum called Ontology SIG, the participants who expressed interest in this group as long as such a group exists should be allowed to go beyond a meta model.
A meta model, to me is the equivalent of an XSD. I helps validation, it checks for the grammar for the language spoken is correct, the statement could very well be to borrow from a poem,
It would be perfectly valid (context awareness, i am reminded of John Strassners post) to say "would like a footlong with everything on it please" in a blockbuster store. So we are back to the killed thread i am afraid, context aware policies and behaviors. But even that belongs in a thread for PBNM or PDNM,
But for a thread that calls itself Ontology SIG should be allowed to explore beyond policies and contexts. even.
Therefore a
a)SID metamodel/Interface Metamodel group can have as a deliverable to come up wth XSD equivalent of Models and Interfaces.for which BTs metamodel doc could be a base. Additional deliverables, automated consistency checks, both static when a new model entities or interfaces are defined(similar to strong typing carrier over to a higher layer here) up to fine grained interface signature/parameter level dynamic checking. We could even extend this to automatic translation between instances of two diferent meta models. About here the name justifies a stop.
b) PBNM which will deal with how to express policies and behaviours an context awareness and determining what are the typical policies that are useful.
c) Onotology SIG, where we go beyond metamodels and context awareness is there a fundametnal structure to knowledge representation, are there fundamental operations on those which would formalize knowledge evolution (collective knowledge, learning). I like Pauls definition, though it could turn out to be ambitious with a narrower focus to our domain by trading off loss in generality, we would have a surer sight of the goal.
Additional deliverables would be,
Exhaustively list concrete Network Semantic Instances.
Map problem space in telecommunication network operations to NEtwork Semantics
Specify using knowledge evolution as above for
   Interface harmonization without loss of semantics
       Simple RCA/SIA
       Self-*
       Trending and Planning.
If the feeling is above is too much to shoot for, i will go with it, but the group should cease to be called Ontology SIG, Ontology and its potential applications to Telecommunications Network
go beyond metamodeling is my feeling.
Sitaraman


Joel Fleck II , Chief Architect , Hewlett-Packard 8/16/2011 5:41:58 PM
Wedge et al:

One of the key directions in research and industry looking forward is the use of a model-driven approach for the construction of business vocabularies, rules, ontologies and technology-neutral representations of the models. In particular, recent research has demonstrated a very usable approach using the OMG’s Semantics for Business Vocabulary and Rules (SBVR) as the basis for transformations to OWL 2.0 (for ontologies and ontological based reasoning –read that semantics), and as the basis for N3 or N-tuple technology-neutral based representations for the policy language(s) of choice, to UML and OCL for the generation of code stubs, and finally to XPATH 2.0 for testing potential interfaces. For reference, I refer to a paper in IEEE Communications August, 2010 “Multidomain IT Architectures for Next-Generation Communications Service Providers”.

Just as an FYI, there has been significant effort on the formulation of an SBVR core for the specification of inter/intra domain relationships and the management of these relationships (and the transformations of that work into UML, OWL, and N3). The methodology has been successfully proven and shown to be very valuable in capturing the business perspective and building software architecture models from them. I would strongly advise investigation into these approaches.

However, at the current time, it would be premature to suggest that the actual models (and a couple of the transformation engines) would be contributed to the TMF as these component parts currently fall into one of two categories; either proprietary work of a consortium, or proprietary work of a individual/corporation.

Best Regards,
Joel

Joel J. Fleck, II
Chief Architect
Head of Standards
HP Software CTO
Hewlett-Packard Company
200 Connell Parkway, Suite 500
Berkeley Heights, NJ 07922

+1 908.898.4091 Phone
+1 908.416.1706 Mobile
+1 908.898.4887 Fax
joel.fleck@hp.com

Kenneth Dilbeck , Director, Technical Program , TM Forum 8/16/2011 5:16:33 PM
Folks,
Anyone familiar with SADL http://sadl.sourceforge.net/?
Perhaps if both groups can use the same tool base, Eclipse, the chances of keeping them together would increase. The JOSIF work is Eclipse based as well.
It would seem having tool support for something as rigorous as an ontology would be helpful. I have only taken a quick look and this may be more oriented towards reasoning systems but thought I would ask.
We do need to make sure there is a very open line of communication between the FAST meta-model team and any other ontology work that takes places to avoid duplication effort and development of conflicting answers.
Regards,
Ken
Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd. 8/16/2011 4:47:53 PM
Hi Wedge,
   Isnt modelling itself an abstraction activity, therefore 2 and 1 are same?
Sitaraman
William Greene , Technical SME; CTR DES-E , DOD 8/16/2011 4:46:08 PM
Classification: UNCLASSIFIED
Caveats: NONE

I’m a little shy of using “automated reasoning” as a goal. I’m personally satisfied with “automated services” as a more reachable goal. Perhaps, "automated, adaptive services” as a stretch goal.

I'm simply happy to start with defining the grammar, syntax, semantics. Also, maybe a few basic patterns that support first understanding Frameworx as a model and then basic operations on information contained in the model.

Eventually, (but not soon ;-) we might get to a point where we can use this to, for example, “frame the set of available interface solutions for domain X and iterate across these to determine an optimal decision tree for function L: federate with the similar decision tree coming from Domain Y.

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

David Milham , Chief Architect , TM Forum 8/16/2011 4:30:55 PM
One needed to be on the call to hear the full background and at the presentation at TAW Baltimore
The key point made by Steve was that this was an example of an enterprise repository solution some aspects of which need to be embedded in Frameworx, Without such a metamodel SP like BT assert that is it difficult to practical evolve systems interfaces business services etc. And it is not primarily a technical issue it is a management and governance need on Frameworx.

Dave
Dave Milham
Chief Architect
t  +44 5601 855561 m+44 7802 902451
edmilham@tmforum.org

From: Exploring New Project Ideas [mailto:newprojects@community.tmforum.org] On Behalf Of Sitaraman Wills
Sent: 16 August 2011 21:14
To: Dave Milham
Subject: [Exploring New Project Ideas] TMF Ontology SIG [#169800]

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.
"David Milham" wrote:
Hi Dave,
Need a favor,

The following document seems to be a MetaModel proposal from Steve Orobeck, this looks like more of a solution document than a list of problems bothering a service provider as you seem to have intended. Am i looking at the right document and if so a pointer to where the problem statements are would help.

If you want to see what is bothering at least one service provider you should look up this contribution.
http://www.tmforum.org/Community/groups/frameworxarchitecture/contributions.aspx?id=artf2509
which was originally presented at TAW Baltimore in the cross Frameworx meetings

The following document gives me a permission denied error, would much aprpeicate if it can be sent as an attachment.
Thanks much.
Sitaraman

http://collab.tmforum.org/integration/viewcvs/viewcvs.cgi/trunk/?root=business_services&system=exsy1002

but we discovered that the individual meta-models don’t line up as we would wish.
Hence we started a top down exercise to complement the prior approach.

What we need now is hand and bodies to work through the patterns we need to use and identify and rectify the gaps in the existing component frameworks.

Dave
Dave Milham
Chief Architect
t +44 5601 855561m+44 7802 902451
edmilham@tmforum.org

From:Information Framework (SID) [mailto:sid@community.tmforum.org] On Behalf Of Sitaraman Wills
Sent: 16 August 2011 19:05
To: Dave Milham
Subject: [Information Framework (SID)] TMF Ontology SIG [#169800]

Sitaraman Wills, Senior Software Engineer , NMSWorks Software Pvt. Ltd.
At the end fo the day, we are TMF first and then Ontologists (to various levels of expertise, from bystanders, through believers, via budding through to been-there-done-thats ).
Perhaps we must have another line of attack from the ServiceProviders side. We could ask operators running their networks day in and day out...tell me in plain english or you native language what are the things that bother you in your job.
There must be complete freedom in responding, no tracking along a direction with a set of 10 questions and the like.
Then we huddle together and extract semantics out of them. Then, we can have that comfortable feeling that, at the end of the day, what we come up with is going to make an important set of people that drive an all pervading industry, smile, courtesy TMF.
Sitaraman
Information Framework (SID)
Topics:General

Shared In:4 Communities
Exploring New Project Ideas
Topics:General

Shared In:4 Communities
Paul Hanke , Sr Professional Staff , Johns Hopkins University Applied Physics Lab 8/16/2011 4:15:48 PM
... I'm not saying that this is what everyone thinks of when they hear the word "ontology", but here is what a deliverable could look like:

An extensible, machine-readable, and foundational logical theory of service provider operations that supports automated reasoning and automated interoperability even across arbitrary domain-specific extensions.

That is, unlike a data model expressed in XML Schema where the best a machine can do with instance data is simply ingore any extensions it is unfamiliar with, an ontology expressed in OWL opens the possibility that a machine can formally manipulate instance data (in a way that mimics human understanding) to include extensions that it has never seen before ... perhaps this is what the FAST team is trying to address with meta-models, I couldn't say ... but even if it is, at this formative stage I see room for both thrust areas to poke and prod at the envelope of the possible.

William Greene , Technical SME; CTR DES-E , DOD 8/16/2011 4:15:42 PM
Classification: UNCLASSIFIED
Caveats: NONE

From Stephen, " Also, the JOSIF open source project has define and implemented a meta-model for interfaces."

Brilliant work!

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

David Milham , Chief Architect , TM Forum 8/16/2011 4:08:59 PM
Steve

Correct thanks for reminding us of the location of that information

Dave
Dave Milham
Chief Architect
t   +44 5601 855561   m+44 7802 902451
edmilham@tmforum.org

From: Exploring New Project Ideas [mailto:newprojects@community.tmforum.org] On Behalf Of Stephen Fratini
Sent: 16 August 2011 20:52
To: Dave Milham
Subject: [Exploring New Project Ideas] TMF Ontology SIG [#169800]

Stephen Fratini , Senior Engineer , Telcordia Technologies
Dave and All,

Also, the JOSIF open source project has define and implemented a meta-model for interfaces.
See the JOSIF wiki pages at
http://sourceforge.net/apps/mediawiki/openoss/index.php?title=Interface_Meta-Model

Steve

---
Stephen Fratini
Telcordia Technologies, Chief Strategy Office
Chair of the TM Forum Integration Program
Home Office: +1 732 962 1500 (I tend to be at this number more often)
Office: +1 732 699 2226
Work Cell: +1 732 853 2207
Email: sfratini@telcordia.com

Exploring New Project Ideas
Topics:General

Shared In:4 Communities
Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd. 8/16/2011 4:06:43 PM
"David Milham" wrote:
Hi Dave,
Need a favor,

The following document seems to be a MetaModel proposal from Steve Orobeck, this looks like more of a solution document than a list of problems bothering a service provider as you seem to have intended. Am i looking at the right document and if so a pointer to where the problem statements are would help.

If you want to see what is bothering at least one service provider you should look up this contribution.
http://www.tmforum.org/Community/groups/frameworxarchitecture/contributions.aspx?id=artf2509
which was originally presented at TAW Baltimore in the cross Frameworx meetings

The following document gives me a permission denied error, would much aprpeicate if it can be sent as an attachment.
Thanks much.
Sitaraman

http://collab.tmforum.org/integration/viewcvs/viewcvs.cgi/trunk/?root=business_services&system=exsy1002

but we discovered that the individual meta-models don’t line up as we would wish.
Hence we started a top down exercise to complement the prior approach.

What we need now is hand and bodies to work through the patterns we need to use and identify and rectify the gaps in the existing component frameworks.

Dave
Dave Milham
Chief Architect
t +44 5601 855561m+44 7802 902451
edmilham@tmforum.org

From:Information Framework (SID) [mailto:sid@community.tmforum.org] On Behalf Of Sitaraman Wills
Sent: 16 August 2011 19:05
To: Dave Milham
Subject: [Information Framework (SID)] TMF Ontology SIG [#169800]

Sitaraman Wills, Senior Software Engineer , NMSWorks Software Pvt. Ltd.
At the end fo the day, we are TMF first and then Ontologists (to various levels of expertise, from bystanders, through believers, via budding through to been-there-done-thats ).
Perhaps we must have another line of attack from the ServiceProviders side. We could ask operators running their networks day in and day out...tell me in plain english or you native language what are the things that bother you in your job.
There must be complete freedom in responding, no tracking along a direction with a set of 10 questions and the like.
Then we huddle together and extract semantics out of them. Then, we can have that comfortable feeling that, at the end of the day, what we come up with is going to make an important set of people that drive an all pervading industry, smile, courtesy TMF.
Sitaraman
Information Framework (SID)
Topics:General

Shared In:4 Communities
William Greene , Technical SME; CTR DES-E , DOD 8/16/2011 4:02:35 PM
Classification: UNCLASSIFIED
Caveats: NONE

I think we here and also the FAST team needs to distinguish between meta-model as:
- a model of a model
- an abstraction of a model.

At least in ontology, it should be an abstraction of a model.

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

Stephen Fratini , Senior Engineer , Telcordia Technologies 8/16/2011 3:40:13 PM
Dave and All,

Also, the JOSIF open source project has define and implemented a meta-model for interfaces.
See the JOSIF wiki pages at
http://sourceforge.net/apps/mediawiki/openoss/index.php?title=Interface_Meta-Model

Steve

- Stephen
William Greene , Technical SME; CTR DES-E , DOD 8/16/2011 2:24:04 PM
Classification: UNCLASSIFIED
Caveats: NONE

Yea, Jerome, not so easy ;-)

These are characterizing statements; perhaps true; but neither providing definitions nor establishing relationships.

For example, Behavior and Lifecycle, must be defined in the semantics of the Ontology. The metamodel executes a ontological syntax, a grammar, in applying behavior and Lifecycle to specific targets. (first pass)

Sitaraman, I think, raises an interesting supposition: is the Frameworx metamodel recursive or self referential? Should that be a structural requirement?

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd. 8/16/2011 1:57:06 PM
At the end fo the day, we are TMF first and then Ontologists (to various levels of expertise, from bystanders, through believers, via budding through to been-there-done-thats ).
Perhaps we must have another line of attack from the ServiceProviders side. We could ask operators running their networks day in and day out...tell me in plain english or you native language what are the things that bother you in your job.
There must be complete freedom in responding, no tracking along a direction with a set of 10 questions and the like.
Then we huddle together and extract semantics out of them. Then, we can have that comfortable feeling that, at the end of the day, what we come up with is going to make an important set of people that drive an all pervading industry, smile, courtesy TMF.
Sitaraman
Jerome Hannebelle , Director, Open Innovation, FT/IMG/Technocentre/BP , France Telecom 8/16/2011 1:49:04 PM
Model experts are certainly more qualified than I, but I would think :
1) While an ontology focuses on giving a meaning(= a semantic) to "objects" (sometimes not as formally as a metamodel), a metamodel probably goes beyond in also looking at behavior, lifecycle, ...
2) The metamodel may define abstract, unifying concepts that may not show up in an ontology (because they are abstract).
Mr experts, please jump in !
Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd. 8/16/2011 1:34:32 PM
I am particularly interested (from the most generatl interface harmonization perspective) in the following goals.
- Can we structure Semantics (we need not consider the general programming language semantics, a perhaps more difficult problem, but can confine to the most frequently occuring, to begin with, semantic patterns in telecommunication networks) Sub goals would be then
       a) What are the frequently occuring "network semantics" both across and within Networks. We can and perphaps must free ourselves of all constraints and feel free to to be abstract here. Then
       b) Can we represent them in COMMON structural pattern, concretely well defined with as little ambiguity as possible.
- Now that we have a COMMON pattern, for the most popularly occuring (to keep things simple at first) concretizations seen above
        a)Represent different instances of the same concrete Semantics in the common pattern and lay them out.
        b)In the abstract (free form) how do we characterize the "extent of difference" between the two.
        c)How do we concretize the differnce expression (procedures, algorithms on the said common structure) that unambigously characterizes the diff. That is, from the diff, one must be able to get to the other from the other, other and converse.
        d) If we do reach c) then hopefully the holy grail of merge (All of A in B, or All of B in A or B = A+B or B = B-A and other possibilities) will be a simpler extension.
       I am hoping, well i would go the extent of saying it is almost a hunch, that the common strcture would have aspects of self-simlarity. Therefore i am hoping the above operations can be recursively applied to whatever turns out to be the "Atomic Operation" on the said structure and therefore be able to do complex "semantic merges", "Semantic Morphs" or Semantc Overlays and such operations out of the recursive operations composed out of the atomic operations.
      If we reach here we have set ourselves up to casting these a)Structure and b)Unit )perations in telecommunicational semnatic processing (due credits to the Chemical Engineering clan),
    Then if only we get lucky and characterize state (a poorer cousin of full semantics perhaps) with the same pattern, if state is a simpler form of semantics it does seem likely,
-Record State of network entities, of groups of network entities (same state compositional types as above, we can reuse those)
- Give said state(s) as input to Unit Operations and its constituents, and based on the input call suitable child operations.... We have context aware, semantic processing,
- Now have state change as part of the child operation and there you have feedback control, with semantics and syntax, a partial learning system, now we can appeal fo more higher levels of concretized math, Stability, Pole-Zero Analysis and other such intriguing things with which Control Theorists work their systems....?
And here it is... I would much appreciate peoples time in responding why that common structure cannot be the simple TREE of arbitray depth and in general a system represented by a forest of such trees? Concrete examples where it fails in achieving a given functionality would help much
Thanks.
Sitaraman

William Greene , Technical SME; CTR DES-E , DOD 8/16/2011 1:23:22 PM
Classification: UNCLASSIFIED
Caveats: NONE

Jerome, could you expound on " believe relates to FAST team's intent to define the Frameworx's Metamodel." My gut feeling is you are right (that is, I am not disputing this comment at all). I feel this is an important statement; but I need some help seeing exactly how the ontology and the metamodel are linked.

From: Wedge Greene

Craig Gallen , Consultant OpenNMS group , The OpenNMS Group, Inc. 8/16/2011 1:08:15 PM
Hi Pierre,
As discussed, I am very interested in this and would like to include the ten University labs I represent through the web Science trust. www.webscience.org
Cheers
Craig
Jerome Hannebelle , Director, Open Innovation, FT/IMG/Technocentre/BP , France Telecom 8/16/2011 1:06:36 PM
Hi Pierre,

This is definitely an interesting topic, which I believe relates to FAST team's intent to define the Frameworx's Metamodel.
Jerome
William Greene , Technical SME; CTR DES-E , DOD 8/16/2011 1:02:58 PM
Classification: UNCLASSIFIED
Caveats: NONE

I'm interested. However, I think after establishing a basic syntax and some inclusive semantic patterns (e.g., search, aggregation, collaboration, policy), fleshing out the various semantic definitions can be handled in the many traditional workgroups. What we eventually need is for each team to feed their semantics into the overall ontology just like we currently feed the SID and eTOM.

My specific interest today is on how an ontology will work with policy.

I do not see we need to get it perfectly right the first time; just we get it close enough to do much of the work... old 80/20 rule.

From: Wedge Greene
* IMHO: Comments are entirely my own and in no way reflect the DoD or anyone I work with.*

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd. 8/16/2011 12:46:06 PM
Hi Pierre,
Meself in it, initially however i request i be granted aceess to discussions and confine myself to observing and learning by participitating in those discussions. I do hope that would be ok.
Thanks.
Sitaraman
Alex Zhdankin , Modeling Architect , Cisco Systems 8/16/2011 12:42:30 PM
Count me in.
Alex.
John Wilmes , Chief Technical Architect, Communications Sector , Progress Software 8/16/2011 12:27:07 PM
I would like to participate especially with reference to SID and eTOM.
John
Paul Hanke , Sr Professional Staff , Johns Hopkins University Applied Physics Lab 8/16/2011 12:27:00 PM
... responding with interest :-)
Bob Natale , Chief Engineer, Joint NetOps & Cyber Defense , MITRE 8/16/2011 12:16:44 PM
Count me in, Pierre.
Pierre Gauthier , CEO and Principal Consultant , Oss Wave 8/16/2011 11:50:18 AM
I would like to see if there is enough interest for creating an Ontology SIG within TMF.
I sugest that the initial focus be on establishing the foundation for a TMF Frameworx Ontology.
It is important that the set of proposed ontologies provides complete coverage of the Frameworx. This involves conceptual modeling of eTOM, SID, TAM but also complete coverage of the Business Services and TIP.
Please reply to that post if you are interested in participating to the TMF Ontology SIG.
End of Discussion History

Full thread available online here: http://www.tmforum.org/community/groups/exploring_new_project_ideas/forum/p/169800/192624.aspx?utm_campaign=community&utm_medium=exploring_new_project_ideas&utm_source=alert#192624

This message was sent to sestes@cisco.com and follows the subscription preferences of your membership of the Exploring New Project Ideas TM Forum Online Community.

To unsubscribe immediately, please click here and click the 'Edit Subscriptions' button from the Email Subscription preferences section located on the right hand side of the page just underneath the menu bar. To cut down on email volume, you may also choose to receive your discussion notifications via a daily or weekly digest email as opposed to immediate separate emails.

TM Forum, 240 Headquarters Plaza, East Tower, 10th Floor, Morristown, NJ 07960-6628, USA. Call +1 973 944 5100. info@tmforum.org

[/collapse]

RE: TMF Ontology SIG [#169800]

VinayD — Wed, 24 Aug 2011 05:10:15 GMT

Yes , I agree with Pierre

Vinay

From: Exploring New Project Ideas [mailto:newprojects@community.tmforum.org] On Behalf Of Pierre Gauthier
Sent: Tuesday, August 23, 2011 10:04 PM
To: Vinay Stephen Devadatta (WT01 - GMT-Telecom Equipment)
Subject: [Exploring New Project Ideas] TMF Ontology SIG [#169800]

Pierre Gauthier , CEO and Principal Consultant , Oss Wave

Hi,

Clearly moving towards the semantic reasoning is the goal.

One of the benefit of establishing a Frameworx Ontology will be to enable the use of the semantic technologies i.e

Ontology Reasoners and at least be able to answers Competency questions on the Frameworx.

Regards Pierre

On 2011-08-22, at 10:05 AM, David Milham wrote:

Exploring New Project Ideas
Topics:General

Shared In:4 Communities

Shared In:Information Framework (SID) : General
Business Process Framework (eTOM) : General
TM Forum Integration Program (TIP) : General
Frameworx Architecture Steering Team (FAST) : General

r>

Full thread available online here:
http://www.tmforum.org/community/groups/information_framework_sid/forum/p/169800/192332.aspx?utm_campaign=community&utm_medium=information_framework
_sid&utm_source=alert#192332

----------------------------------------------------------------------------------------------------------------------------------------------------
This message was sent to dingersoll@versant.com and follows the subscription preferences of your membership of the Information Framework (SID) TM
Forum Online Community.

To unsubscribe immediately, please click here and click the 'Edit Subscriptions' button from the Email Subscription preferences section located on
the right hand side of the page just underneath the menu bar. To cut down on email volume, you may also choose to receive your discussion
notifications via a daily or weekly digest email as opposed to immediate separate emails.

TM Forum, 240 Headquarters Plaza, East Tower, 10th Floor, Morristown, NJ 07960-6628, USA. Call +1 973 944 5100. info@tmforum.org

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.

8/16/2011 5:54:44 PM

Ah!

Sitaraman

Stephen Estes , Business Process Architect , Cisco Systems

8/16/2011 5:48:11 PM

http://protege.stanford.edu/

-- Stephen

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.

8/16/2011 5:46:19 PM

Sitaraman

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.

8/16/2011 5:42:11 PM

Dave/Jerome,

Yes, I would go with any set of deliverable as long as it is useful to the industry and those are clearly defined.

Having said that, i do feel that in a forum called Ontology SIG, the participants who expressed interest in this group as long as such a group exists should be allowed to go beyond a meta model.

A meta model, to me is the equivalent of an XSD. I helps validation, it checks for the grammar for the language spoken is correct, the statement could very well be to borrow from a poem,

But for a thread that calls itself Ontology SIG should be allowed to explore beyond policies and contexts. even.

Therefore a

b) PBNM which will deal with how to express policies and behaviours an context awareness and determining what are the typical policies that are useful.

Additional deliverables would be,

Exhaustively list concrete Network Semantic Instances.

Map problem space in telecommunication network operations to NEtwork Semantics

Specify using knowledge evolution as above for

Interface harmonization without loss of semantics

Simple RCA/SIA

Self-*

Trending and Planning.

Sitaraman

Joel Fleck II , Chief Architect , Hewlett-Packard

8/16/2011 5:41:58 PM

Wedge et al:

Best Regards,

Joel

Joel J. Fleck, II
Chief Architect

Head of Standards

HP Software CTO
Hewlett-Packard Company
200 Connell Parkway, Suite 500

Berkeley Heights, NJ 07922

+1 908.898.4091 Phone
+1 908.416.1706 Mobile
+1 908.898.4887 Fax
joel.fleck@hp.com

Kenneth Dilbeck , Director, Technical Program , TM Forum

8/16/2011 5:16:33 PM

Folks,

Anyone familiar with SADL http://sadl.sourceforge.net/?

Perhaps if both groups can use the same tool base, Eclipse, the chances of keeping them together would increase. The JOSIF work is Eclipse based as well.

Regards,

Ken

Sitaraman Wills , Senior Software Engineer , NMSWorks Software Pvt. Ltd.

8/16/2011 4:47:53 PM

Hi Wedge,

Isnt modelling itself an abstraction activity, therefore 2 and 1 are same?

Sitaraman

William Greene , Technical SME; CTR DES-E , DOD

8/16/2011 4:46:08 PM

David Milham , Chief Architect , TM Forum

8/16/2011 4:30:55 PM

One needed to be on the call to hear the full background and at the presentation at TAW Baltimore

Dave

Dave Milham

Chief Architect

t +44 5601 855561m+44 7802 902451
edmilham@tmforum.org

From:Exploring New Project Ideas [mailto:newprojects@community.tmforum.org] On Behalf Of Sitaraman Wills
Sent: 16 August 2011 21:14
To: Dave Milham
Subject: [Exploring New Project Ideas] TMF Ontology SIG [#169800]

Sitaraman Wills, Senior Software Engineer , NMSWorks Software Pvt. Ltd.

"David Milham" wrote:

Hi Dave,

Need a favor,

If you want to see what is bothering at least one service provider you should look up this contribution.

http://www.tmforum.org/Community/groups/frameworxarchitecture/contributions.aspx?id=artf2509

which was originally presented at TAW Baltimore in the cross Frameworx meetings

The following document gives me a permission denied error, would much aprpeicate if it can be sent as an attachment.

Thanks much.

Sitaraman

http://collab.tmforum.org/integration/viewcvs/viewcvs.cgi/trunk/?root=business_services&system=exsy1002

but we discovered that the individual meta-models don’t line up as we would wish.

Hence we started a top down exercise to complement the prior approach.

What we need now is hand and bodies to work through the patterns we need to use and identify and rectify the gaps in the existing component frameworks.

Dave

Dave Milham

Chief Architect

t +44 5601 855561m+44 7802 902451
edmilham@tmforum.org

Sitaraman Wills, Senior Software Engineer , NMSWorks Software Pvt. Ltd.

At the end fo the day, we are TMF first and then Ontologists (to various levels of expertise, from bystanders, through believers, via budding through to been-there-done-thats ).

There must be complete freedom in responding, no tracking along a direction with a set of 10 questions and the like.

Sitaraman

Information Framework (SID)
Topics:General

Shared In:4 Communities

Exploring New Project Ideas
Topics:General

Shared In:4 Communities

Paul Hanke , Sr Professional Staff , Johns Hopkins University Applied Physics Lab

8/16/2011 4:15:48 PM

... I'm not saying that this is what everyone thinks of when they hear the word "ontology", but here is what a deliverable could look like:

William Greene , Technical SME; CTR DES-E , DOD

8/16/2011 4:15:42 PM

David Milham , Chief Architect , TM Forum

8/16/2011 4:08:59 PM

Steve

Correct thanks for reminding us of the location of that information

Dave

Dave Milham

Chief Architect

t +44 5601 855561 m+44 7802 902451
edmilham@tmforum.org

From:Exploring New Project Ideas [mailto:newprojects@community.tmforum.org] On Behalf Of Stephen Fratini
Sent: 16 August 2011 20:52
To: Dave Milham
Subject: [Exploring New Project Ideas] TMF Ontology SIG [#169800]

Stephen Fratini, Senior Engineer , Telcordia Technologies

Dave and All,

Also, the JOSIF open source project has define and implemented a meta-model for interfaces.

See the JOSIF wiki pages at

http://sourceforge.net/apps/mediawiki/openoss/index.php?title=Interface_Meta-Model

Steve

---

Stephen Fratini

Telcordia Technologies, Chief Strategy Office

Chair of the TM Forum Integration Program

Home Office: +1 732 962 1500 (I tend to be at this number more often)

Office: +1 732 699 2226

Work Cell: +1 732 853 2207

Email: sfratini@telcordia.com