BALTIMORE — MEF16 — MEF is working hard to reinvent itself, intending to move well beyond Carrier Ethernet to become the forum that enables end-to-end service orchestration across multiple carrier networks in the virtualization era. The man driving this change, CTO Pascal Menezes, says MEF will become “very scrappy and very agile” and bring on board its own code development resources as well. (See MEF Redefining Its Role.)
In his keynote address Wednesday and in an interview with Light Reading, Menezes stressed the fact that MEF isn’t becoming an open source group but will have open source elements to its work. Nor will MEF focus on developing code, though it will create code to enable its specifications in an effort that will soon go well beyond the hackathon events it has been staging at its annual conferences, in support of its Lifecycle Service Orchestration (LSO) project.
MEF is also taking the lead in working with other groups, mostly open source organizations but also the TM Forum . It will work with these groups to knit together existing resources and code, enabling things such as open applications programming interfaces within carrier operations (northbound-southbound) and between carriers (eastbound-westbound). (See MEF Mind Meld Unites Groups on NaaS.)
The idea is to leverage the power of the $80 billion Carrier Ethernet market that MEF helped enable in creating business networks that are more agile than Ethernet and can be delivered on-demand via IP, including point services such as security and more. MEF calls this the Third Network and at its heart is the Lifecycle Services Orchestration effort it is creating to link together multi-carrier networks. (See MEF Launches Lifecycle Service Orchestration Report.)
“We have to be able to connect providers in one big federated network,” Menezes tells Light Reading. “So when we automate services and allow changes to happen on a machine-to-machine basis, we have the language that is used to communicate between providers,” he says. “We know we need to automate, to get people out of the way — we need APIs and we need them now.”
There is growing energy behind the idea of combining open source efforts with more traditional standards or specifications in a way that makes sense for operators but doesn’t slow the entire virtualization process down. (See Defining MANO: Open Source vs. Standards and AT&T: MEF Could Catalyze Key Specs.)
The old MEF way of doing things would be to launch a major specifications project, define everything it needs to cover and work it through committees and detailed processes for up to three years before releasing a specification. The new agile methodology still creates a list of what needs to be done — based on input from its service provider members — but then prioritizes that and breaks it down into small chunks of defined use cases.
“What is important is for us to get to small-chunk-sized use cases that we can describe and model them and then prove them out in code in our open initiatives and sprints, and this is the agile model,” Menezes said in his MEF16 keynote. “You take the things you want to do, put them in a backlog, prioritize the first few, go do a sprint on it, deliver on it, have your stakeholders say if this is what they want, then iterate again. This is going to be our new process on how we are delivering as the MEF, we are going to be a very scrappy, agile organization.”
Keep reading on Light Reading.