I recently had the pleasure of conducting a one-on-one interview with Michael Sullivan-Trainor, Executive Analyst with Technology Business Research in New Hampshire. Michael is a thought leader in the NFV space and I quizzed him about the business case that might be underlying the technology hype around NFV. This is the first of a two-part series.
Q. All the excitement is around NFV and getting it up and running, why haven’t we looked more deeply into the business case?
A. The attention has been primarily on revenue generation, with standardization bringing efficiency to the network transformations and finding what services we can drive end-to-end in the digital services space. We’ve been documenting business cases for that. The trigger here is the industry specifications group (ISG) affiliated with ETSI that basically said there is a cost savings value with the implementation of NFV and we really need to tie down the specs and put it towards use cases where we drive some cost savings.
Operators struggling with their margins are saying – “we’ve got the revenue initiatives going now but what can we do in the cost reduction side?” That’s where the business case initially lies. I think there are breadcrumbs that lead us from cost reduction to revenue agility as people are looking at the long held goal of speed on new service innovation so it aligns with the path towards new service innovation, but begins with cost savings.
Q. Where do you see these cost reductions coming from? Just introducing NFV on its own is not going to be a great cost saver initially, probably the opposite, but where do you see it being a real cost saver?
A. It’s really about scale. The industry approach initially took it literally and said let’s virtualize functions but I don’t think that by virtualizing any individual function much is to be gained. Something incremental you do by, say, virtualizing policy control, for example, and that enables the business case for policy control to be more easily generated as a distributed technology across different layers of the network. You’ve done something there that perhaps triggers better policy control and better quality of service initiatives by just virtualizing that function and making it able to be distributed more cost effectively. So there is a small kind of business case, but the real value is what the operators have begun saying occurs on a network scale, taking the concept of programmable network by virtue of implementing a software-defined network (SDN).
But there’s a great deal of programmability automation that comes with the orchestration promised in NFV. So think about more and more programmability being hooked up across the network to basically take out a lot of the manual OPEX requirements as well as some of the CAPEX and OPEX related to pure hardware, but it isn’t very interesting just doing it for one function here or there, it has to be done on a network scale. I think that’s where we are headed and that’s where the value comes from.
Q. What will an operator have to do to roll out NFV? Is it a transformation project, will they have to add new components, is it only software driven? And what is the cost implication?
A. It’s a transformation project, but that said, the observation is the place to start guided by the Industry Specification Group (ISG) use cases. So then you are able to understand all those components and build a business case. In a use case you can see the revenue you’re impacting whether it’s increasing value or not. You can see which costs you are taking out and which costs you are putting in.
So obviously there’s a significant upfront professional services cost for consulting on how to transform the use case into a new network architecture. Then the design of the software platform is an implementation that is a net new. However, the opportunity in that same use case to replace things on an operational basis, primarily you may replace some of the hardware but you’re definitely going to invest in the software. So the total cost ownership of NFV enabled use case versus a traditional use case is about the savings on the operational costs after the initial implementation costs are going to be very important.
Q. Is that a challenge do you think for people within the operation pushing management to spend money on NFV?
A. Well, I think that’s something that comes out of the CTO’s office. The CTO is always tasked with what can we do next and better and so it will strike the operational side of the organization either positively or negatively. Positively if they are saying yes we see all these new things we need to do and we are shorthanded and we really have to pull costs out of the system and we need more automation. Negatively if there are all these people doing manual functions and they’re quite happy, thank you. What happens to them? How do they get retrained? So it does vary where the operator is sitting right now. Sometimes I think the newer operators that have less invested in human capital might be faster to implement this on a broader scale.
Watch out for the second part of this interview where we investigate use cases, supplier involvement and who the main players in NFV are.