How a New Alliance Could Drive NFV into the Real World

One of the things that came out of Light Reading’s NFV Everywhere event was a sense that the critical MANO element of NFV has somehow lost its luster.  There have been some articles on that topic, and one of the most active members of the NFV ISG has posted a blog as well.  IMHO, the truth lies somewhere between, and interestingly we have an announcement to report on that may weigh in and decide the issue.

MANO is the true innovation of NFV.  The basic notion is that a service is constructed by somehow coercing cooperative behavior from a set of resources, a process called “orchestration” (and the “O” in MANO).  The orchestration also coordinates the management of both services and resources (hence the “MAN”).  At this high level, MANO deserves all the pundits you can heap on it.

At another level, MANO fell short, or rather the ISG did.  From the very first, it should have been obvious that you cannot orchestrate services and not orchestrate management, that virtualization is either universal or it’s been nailed to the ground somewhere.  Operations and management processes from both the NMS/NOC and OSS/BSS world needed to be involved, and these were declared out of scope.  That was a critically bad move.

Even the operators who launched NFV in the first place had abandoned the notion that NFV could be driven by capex benefits alone within a year of their initial white paper.  That left “operations efficiency” and “service agility” as the benefits, and operators knew in the fall of 2013 that these benefits could arise only from automation of the entire service lifecycle.  That means not only integrating NFV management and operations with orchestration from the OSS/BSS down, it means integrating legacy equipment as well.  And remember that both legacy equipment and operations were out of scope to NFV.

But is MANO irrelevant, and if so why?  There are things about MANO as it has evolved that I don’t agree with, but none of it matters because the scope issue is absolutely fatal.  Everything happening now with MANO is really a response to the fact that NFV can’t make its own business case, which means something outside it has to take over.  As for MANO itself, well, since it can’t go “up” to the OSS/BSS level, it’s effectively contracting to become nothing more than a kind of hazy partner to OpenStack.

OpenStack reflects its own good-and-bad.  Rightfully, OpenStack is just an implementation of the NFV Virtual Infrastructure Manager (VIM).  It’s available, popular, and it seems to “orchestrate”, and if you believe the ISG that operations is out of scope to NFV, then it’s sufficient.  That’s resulted in a lot of vendors jumping on OpenStack as being MANO, and that’s helped obscure the fact that something even more than “classic ISG MANO” is needed.  It is needed, because without MANO on a very large scale, larger than the ISG paints it, there will be no NFV business case except for very specialized services and operators.

This is where the announcement comes in.  Overture Networks announced the “integration of Ensemble Service Orchestrator (ESO) with Wind River® Titanium Server, providing carrier-class NFV management and orchestration on the Titanium Server virtualized network infrastructure solution.”  To make this happen, Overture has stepped up from what was initially a very limited vision of NFV, hosted on their own custom carrier-Ethernet CPE, to become a real NFV player.  We thus have the best platform available for NFV hosting in Wind River, combined with one of the very few NFV implementations that can actually make a business case.  The combination will presumably be available through both the Overture Harmony and Wind River Titanium Cloud ecosystems.

What makes this significant is that it’s a kind of populist-NFV-in-a-brown-paper-bag.  Other vendors who don’t have a real NFV implementation (vendors other than Alcatel-Lucent, HP, Oracle, and Overture, and probably Ciena and Huawei) can now get the components to supplement whatever they do have.  With the right documentation and approach, they can turn NFV PoCs and trials into field trials and deployments.  The other NFV solutions are tied to major vendors with their own agendas, which makes these less attractive to the NFV masses.  Other computer vendors (who don’t want to pull in HP), other network equipment vendors (who don’t want Alcatel-Lucent or Huawei), and OSS/BSS players who won’t like any of the other NFV solutions, now have somebody to partner with.

I don’t think this is an indication that Wind River and Intel are ready to field their own NFV approach (though I’d love for them to).  Wind River has major partners like HP with strong NFV stories of their own, and it makes sense for Wind River and Intel to hedge their bets…sort of.  This might be a classic arms-merchant strategy—start a war and arm all sides.  Intel’s chips are likely winners no matter who does NFV, but they could lose if nobody does.  The Overture NFV approach is pretty portable; any server or software vendor could adopt it, which is good for Intel.

There’s still work to do here, though.  Even with Overture’s software there is still a significant integration task associated with making NFV work in a large-scale carrier application.  Operations integration and legacy devices will have to be specialized to the needs of the operator.  I don’t think Wind River is signing on for that, so the next question will be whether Overture can attract some big integrator interest, or maybe a buyer for the company itself.

At the least, this should elevate the NFV dialog.  The Light Reading event reinforces my own discussions with operators; they’re looking for a shift to a business-case-driven approach.  Even those vendors who could make an NFV business case have been silent on what would be involved, and thus unable to prove they could do it.  Silence now is going to be the opposite of golden, because a lot of new players could grab the Overture/Wind River combination, sing pretty, and become a player.  Everyone who, technically speaking, already is a player will have to sing too, and all the singing should finally expose the reality of what is needed for NFV.

Any way this shakes out will be interesting.  NFV needs a business case, and Overture cold make the NFV business case, or at least they could articulate all the required pieces.  They have the Masergy deal in the bag as a proof point for their capabilities.  That’s a lot of credentials for a small player, and you can see there could be a lot of interesting outcomes.