Lately we have
been seeing a huge spike in interest for a variety of applications that
manage large numbers of widely distributed intelligent devices in a
variety of new broadband networks. One of our partners, Accedian
recently put a Nakina application into a US wireless carrier to manage
turn-up and test for a large rollout of Ethernet NIDs in a backhaul
application; in an another, a top equipment vendor has begun
implementing a Nakina Resource Optimization (parameter management)
solution for a national IMS rollout supporting consumer VoIP.
There are a lot of other cases that can’t yet be disclosed publicly,
but they all involve management of networks that are undergoing various
stages of disaggregation. In the consumer VoIP example, over
20 individual classes of network element (media gateway controllers,
routers, access managers, etc.), deployed in multiple instances,
replace a single centralized switch. The new architecture is vastly
more flexible, and takes advantage of the inherent efficiencies of
packet-based transport, but the flexibility and efficiency come at a
cost in terms of increased management complexity. Rev levels, patches,
parameter settings, backups, and security in the disaggregated
environment can’t be managed without new infrastructure and new
methods, and these are often improvised at rollout rather than being
baked into the plan. 
( photo credit: http://www.computerhistory.org/)
It strikes me that what we’re seeing among service providers — and
in the management systems that support them — has a lot in common with
the evolution of enterprise computer and network architectures — in
fact, it’s nothing more than a delayed reflection, played out in an
industry that has vastly longer investment cycles and vastly slower
technical evolution. The movement out of the 1970s mainframe-centric
world and into the ‘peer-to-peer’ minicomputer networking world of the
1980s (the origin of the internet) and then further into the evolution
of ubiquitous computing in the 90s and 00s, gave rise to a whole new
multibillion dollar industry devoted to management support — network
management, PC desktop management, server system administration, and so
forth. It created an opportunity for rapid development of hundreds of
companies, many of which became billion dollar plus players (CA, IBM’s
Tivoli, and BMC for example).
We’re still in the early days of a massive transformation from a
‘mainframe’ era of telecommunications into a disaggregated era — based
on IMS, LTE, IPTV, femtocells, and ethernet transport (to name only a
few examples), with content and service originating from millions of
endpoints around the network, not radiating out from its center. It’s
time for a new ‘operations software foundry’ to start forging the tools
and building the machines that will empower that transformation.