TM Forum Sees Catalyst Role in NFV
SAN JOSE -- Digital Disruption 2013 -- Following an all-day meeting this week with the network functions virtualization (NFV) Industry Specification Group (ISG), the TM Forum is expecting to generate multiple NFV-related Catalyst projects for its Management World event in Nice next June. Those projects would then also serve as proofs-of-concept for the NFV ISG and tackle some of the known management and network orchestration issues.
Catalyst projects are short-term collaborations among TM Forum 's members, including service providers, hardware/software vendors, and sometimes enterprises, which tackle specific problems and result in demonstrations at the Forum's events. Based on member feedback from Monday's event, Ken Dilbeck, VP-strategic programs for the Forum, expects up to three of the Catalysts in Nice will be NFV-related and those will include an additional element.
"We want to set an extra goal for our Catalysts, to construct themselves so they can be submitted as proofs-of-concept for [the NFV ISG] as well," Dilbeck says in an interview here. "We believe that is more constructive than having two separate processes. We also believe that the management concerns around NFV are an area we can focus on with our experience and bring some real progress."
Catalyst proposals are due December 6 and the selection process is in mid-January.
There are multiple reasons why Forum officials think their Catalyst process is a way to tackle some of the network management and orchestration challenges facing NFV. For one thing, Catalyst projects are known for tackling technology challenges ahead of standards, Dilbeck says, and working through issues that can ultimately influence standards in some cases.
"Taking an incomplete specification and doing something with it is not foreign to us -- people [who] want to have a hard specification before they invest aren't used to that," he says.
The Catalyst process also enables fast failure -- the concept of figuring out quickly what doesn't work and moving on, Dilbeck says. And the Forum's process isn't that different from the proof-of-concepts process the NFV ISG is using -- it's more formal in that there are established elements for intellectual property rights and for marketing the projects.
Until this week, contact has been limited between the TM Forum, with its focus on network management, BSS, and OSS, and the NFV ISG, formed in the fall of 2012 under European Telecommunications Standards Institute (ETSI) 's auspices. The Forum had access to the public documents the NFV ISG had released and had some conversations earlier this year at a meeting in Baltimore, but there hadn't been an in-depth discussion.
The California meeting was a bit of serendipity, Dilbeck admits, because it turned out both organizations had separately planned major meetings only 12 miles apart the same week. Monday's all-day workshop was the result of that happy accident and, while there were moments of contention and skepticism on both sides about some things, Dilbeck believes it produced constructive insights for the TM Forum. (See A Big Week for NFV and Will TM Forum Disclose Role in SDN/NFV?)
"The intent was to open up an exchange of ideas," he says. "The ETSI group intends to be done with its work in a year. We wanted to know where it makes sense to hand off certain work items, and to show them what our capabilities are and where we think we have assets that would be pertinent to their work."
Some of Monday's discussion was around what specific interfaces or places in the NFV ISG architecture would most likely engage Forum members. The most likely areas were the interface between OSS/BSS systems and the NFV management and orchestration, the interfaces between the NFV orchestration and lifecycle manager and both element management systems and virtual infrastructure managers, and the interface between the NFV orchestration and lifecycle manager and the virtual infrastructure managers.
Who does what?
But there is no certainty as yet on these issues -- as multiple participants in Monday's session noted, even within individual companies there isn't complete agreement on how the NFV process and the network management and orchestration pieces come together. (See Timeline for NFV Still Hazy.)
Everyone agrees there is much work to do to both create management and orchestration processes and systems for NFV and to determine how and where to link the new stuff to the existing OSSs/BSSs, Dilbeck admits. He believes much of that work can be accomplished within the Forum and its Catalyst project process in an efficient way that avoids vendor-specific or de facto solutions.
Efficiency will be a major issue -- the one other certainly expressed by service providers engaged in the NFV process is that they are moving ahead as fast as they can in adopting this technology, in some cases pulling vendors along with them, and looking for solutions they can adopt now rather than later.
— Carol Wilson, Editor-at-Large, Light Reading