It’s been an interesting week for the markets, in all of the dimensions that drive us forward. There are glimmers of technology change in SDN, there are signs of vendor shifts, and there are macro indicators that might tell us a bit about demand. So, given that it’s Friday and a good day to take a broader look at things, I’ll talk about all three.
NTT’s strategy linking SDN technology to cloud service delivery for hybrid cloud and even cloudshifting applications is potentially interesting. If you look at the infrastructure dimension of SDN, you see it focused primarily on narrow data center switching applications or (in Google’s case) a fairly specialized IP core. The big opportunity for infrastructure SDN is metro, and we’ve been light on examples of things that you could do with it in the metro space. NTT proposes to have data center connections made via an SDN-controlled gateway, which could not only create a framework for metro cloud services to enterprises, it could be an element in NFV implementation.
The data center is the foundation of NFV, but the biggest (and perhaps hardest) step in NFV infrastructure evolution is creating a web of resource connections throughout the metro area. Applications like IMS/EPC and CDN, both NFV use cases, demand interconnection of metro resources. Most of the advanced mobile/behavioral opportunities that operators have identified—both in the enterprise space and for consumers—demand a highly elastic resource pool with good connectivity that lives underneath the orchestration and operationalization layers. Thus NTT may be driving some activity in a very critical area.
On the vendor side, we can see that the spin out of Cisco’s Live event on Insieme does seem to be focusing on something more “operationalizing” in terms of benefits, which I think is both good for the market and smart for Cisco. The goal of SDN must be the support of the evolution of that grand fuzzy artifact we call “the cloud”. If SDN doesn’t do that, it’s just tweaking current processes and won’t matter much in the long run. The challenge in supporting the cloud isn’t just a matter of speed-matching agile application-layer evolution with seven-year-depreciation infrastructure, it includes the question of how you can make all that dynamism work without creating an explosion in operating costs. That problem is so critical to carriers that it swamps the question of how you optimize deployment or whether you even use SDN technology.
We need a new operations model for the cloud, and in fact bodies like the TMF have recognized that for at least four or five years. The problem is that we’ve been unable to create one using the traditional standards processes, largely because those processes get so tied up in evolution they forget the notion of destination. If Insieme can be used to take a bold leap into the operational future and jump over all the dinosaurs of past processes, then it gives us the destination of cloud operations. Getting to that goal can then be addressed, and I know for sure that there are models of operations evolution that can manage both the goal and the route. I’ve presented them to carriers, in fact. What has to happen now is that they get productized.
The first of our broad-market points is the Accenture quarterly report, which was a revenue miss due to shortfalls not so much in outsourcing as in consulting activities. Professional services have been a bright spot for a lot of companies, not the least being IBM. In networking, Alcatel-Lucent, Ericsson, and NSN are increasingly dependent on it. So the question at this point is “Are tech buyers being asked to spend more on professional services to competitively pressured equipment price reductions?” It appears to me based on the market data and my spring survey results (now almost all in from the responders) that buyers think so. Enterprises are starting to adopt that suspicion-of-my-vendor mindset that evolved over the last five years in the carrier space and has substantially poisoned the credibility of the vendors.
The challenge this poses for vendors is multi-faceted. On the surface, having your buyers distrust you has to be a bad thing, and that’s especially true when they’re trying to make a major technology transition. In fact, the poisoning of the credibility well is so serious a problem for vendors that the only reason it’s not hurting them is that nearly all of them have it. Cisco seems to be the vendor who has escaped the credibility slump the best, and not surprisingly it’s the network vendor doing the best in the marketplace. But there’s a deeper point too, and that’s the fact that buyers really do need professional services to help them along, and if they don’t get them because the services are overpriced or the sources aren’t trusted, then major new applications of networking and IT can’t be deployed. That would disconnect all our wonderful new technology options—cloud, SDN, and NFV—from the improve-the-benefit-case side of the process, focusing them entirely on cost reduction. That, as I’ve said for years now, will take networking and tech to a place where few of us in the industry will want to see.
The good news is that I think we’re starting to see, in the “SDN wars” the beginning of a synthesis of a value proposition. Yes, it would have been better had someone simply stood up and announced the right answer, meaning positioned their assets to address the real market value of SDN. However, we’re finding the right spot by filling in a million blank spaces, some substantially useless, and then looking at the impact of each new solution. Eventually, we’ll see the elephant behind the forest of boulders, snakes, and trees.