NFV (Network functions virtualization)

Analyst: Transformation Must Be Top-Down

Transforming telecom network organizations to deliver new services quicker is a top-down exercise, but most of the current efforts are starting at the bottom and working their way up, argues Tom Nolle, a veteran industry consultant and the CEO and president of CIMI Corp. .

Nolle, who launched the CloudNFV initiative and was its chief architect until he stepped down in February, is now developing an open architecture for orchestration and management designed to deliver the service agility telecom network operators say they want, by starting at the top at the services level and working down to the resource level. (See New Group Ties NFV to the Cloud and Top 5 NFV Movers & Shakers.)

That is the opposite approach to what industry standards groups are taking today, Nolle wrote in a blog post on his site today. Groups such as the European Telecommunications Standards Institute (ETSI) NFV Industry Specifications Group, the Open Networking Foundation , and the TM Forum are not setting out to come up with a complete state of orchestration for next-gen services. Like traditional standards group, they have a tendency to try to build up from the bottom.

"If you look at the classic example of OpenFlow, here we are defining a way to control forwarding tables, and we don't have any kind of a management architecture or any kind of service model -- none of the stuff you would expect to have first," Nolle told us. "We haven't even gotten to it yet, because we are all down in the details."

He plans develop such an open architecture for orchestration and management and to offer it to the industry through a series of tutorials he'll develop by late summer. There won't be a charge for accessing those tutorials or using the ideas -- though a vendor that wants specific help will have to pay consulting fees, and the use of Nolle's terminology will require his scrutiny.

Since posting his thoughts on his blog today, Nolle has already had responses from one standards group and a few operators. His goal is to enable a more rapid transition to virtualization in the form of SDN and NFV, as well as to cloud-based systems.

"If I've got three revolutionary technologies and a trillion dollars' worth of legacy infrastructure that I'm going to evolve out of, and I don't have any high-level strategy for putting them together, then how am I ever going to be efficient?" Nolle asked. "I'll be lucky if I can hold my costs to what they are now. I just don't see how we are going to get anywhere with this if we don't take a broader view of this, if we don't take a more holistic view."

— Carol Wilson, Editor-at-Large, Light Reading

Page 1 / 2   >   >>
Liz Greenberg 4/3/2014 | 6:19:18 PM
Re: Is It Possible? @ seven,  I see your point.  I was thinking more from a holistic, integrated architectural approach which would take organizational transformation as well. The issue is that the multiple architectures that you refer to do evolve somewhat independently from each other.  I have even experienced at the expense of each other when organization battle for headcount, dollars, etc.

This is where the CEO does not say "SDN/NFV" but rather something like...we need to transform our company to be more nimble and to do that we need to look at the long view, etc. etc. etc. The higher level management needs to insure that the organization moves together not against each other to achieve the goals and have ways of assessing whether or not it is happening - metrics! 
TaraSeals 4/3/2014 | 4:33:02 PM
Re: Is It Possible? Ha! Yes. Maybe we can come up with a special plaque or something, at the very least. :-) 

It does sound like the holy grail--and even better if it moves service agility and innovation along in a way that boosts economic outlooks for businesses and makes consumer lives that much better. All too often it seems like carriers miss that piece of things because they get bogged down in the middle of the chain.

Maybe I should get a T-shirt made up that says "TEAM TOM!"
brookseven 4/2/2014 | 11:20:12 AM
Re: Is It Possible? Liz,

I think top down means something very different than pretty much everybody else here.

To me, a Top Level Executive (say a CEO) at a major Telco should not be involved deeply in architecture decisions.  What they should be doing setting a direction for how the company has to change in order to grow.  This should drive aligned behavior below them to meet those objectives.  The goal of an executive team is to set a direction and then ensure that their staff sets and executes plans to meet the direction.

Of course, when you have 100K employees keeping that all moving in a direction is a tough challenge.  So I would not expect a CEO to say "We need SDN/NFV".  What I might expect them to say (and even this might be a stretch) "I want to compete on the basis of service agility.  We need to set plans on how we are going to get there."  More likely I would expect something like:  "I want to improve service margins.  What are the best ways that we can differentiate ourselves?"

Then one would have to look at all the various differentiation options and pick some.  From there, goals and metrics can be derived as a plan is established.

One choice for differentiation/change drivers would be a large scale architecture.  My experience in the carrier environment is that it has many architectures that do many things and they evolve somwhat indpendently of each other both in capability and timeframe.   So for example, if I look at SDN then I have to ask the question - "How many organizational layers does an SDN implementation have to cross to build an interesting new service?"  If it is more than 2, then I nod and assume that this will not happen fast or at all.  NFV has the advantage that it can be just a 1 for 1 replacement of existing functionality.  I don't think most of the proponents view it that way and so probably see it evolving differently than I do.


TomNolle 4/1/2014 | 7:08:56 PM
Re: Is It Possible? I'm happy to work with the TMF; my debt to GB942 has been acknowledged many times in the past, and I'm a member.  I can do a private presentation to the TMF leadership on the model in May if you're interested.
kdilbeck570 4/1/2014 | 6:15:36 PM
Re: Is It Possible? Carol, Tom,

I agree with Tom's point about the urgency to explore the requirements of orchestration to meet all the requirements of NFV and a completely vistrualized infrastructure. TM Forum actually has two catalyst targeted for TM Forum Live! in June of this year that are exploring these issues. More input and thought is required. The ZOOM project at the Forum would welcome ideas and inputs on how to expite this work.

TomNolle 4/1/2014 | 3:52:22 PM
Re: Is It Possible? I'm going to give Carol an exclusive when I can talk about this more, and everyone will then be able to judge for themselves (which is what I'd do if I were in your shoes!)  I don't want to dribble stuff out, though; I want a complete architecture that will answer everyone's questions.
mendyk 4/1/2014 | 3:24:54 PM
Re: Is It Possible? Sounds like a great idea, and a great challenge.
TomNolle 4/1/2014 | 3:18:56 PM
Re: Is It Possible? I agree completely.  The right approach IMHO is to define a way to model services that works for both legacy and next-gen elements equally well.  Think of it as being something like a general "infrastructure handler" with a plugin that fits NFV, one for OpenDaylight, one for Cisco NMS provisioning, etc.  You manipulate the models and the infrastructure follows.  But I know it's a lot to expect, and the proof will come when I can get the tutorial done!  But I first have to run things by the operators and key standards bodies to get their input.
mendyk 4/1/2014 | 2:18:24 PM
Re: Is It Possible? Nice idea, but it does run against the firmly established nature of "transformation" (i.e., gradual change to existing reality rather than whole-cloth reinvention). Grand pronouncements from on high tend not to get very far.
TomNolle 4/1/2014 | 1:40:50 PM
Re: Is It Possible? My favorite quote on your topic is from Stevens in 'Computer Networks' where he says "The wonderful thing about standards is that there are so many to choose from!"  I'm not proposing to standardize something as much as define an architecture that leverages the standards we already have.  I'm trying to work through a tutorial on this ASAP, but I want to give operators and the standards bodies a chance to see it first, so it likely won't be until June that I can release something, assuming I don't make changes based on input from these people.
Page 1 / 2   >   >>
Sign In