Well, if you thought that a telecom spending spree or an enterprise market flush with unexpected cash were going to drive Cisco and the industry to unprecedented growth, Cisco’s call likely disabused you of that notion. While Cisco beat most estimates very slightly, its guidance was considered tepid and certainly the continued references to “challenging macroeconomic environment” on the call didn’t boost confidence.
Telecom spending is certainly ramping up a bit, and mobile is the biggest focus of that incremental spending and Cisco’s gains in mobile infrastructure are consistent with the market reality that you invest where you get the biggest return. Enterprise spending is down a bit, and that’s consistent with the fact that networking has become a cost center and not an opportunity at the enterprise level. Cisco’s success with UCS was impressive, but as I said yesterday it’s not enough to pick the low apples of server opportunity here. Cisco needs to drive the bus to the future, not ride along as a passenger.
An area I think is a missed-so-far opportunity is metro. Yes, it’s true that mobile gets the lion’s share of investment but mobile investment is transforming metro networking. There are at least twenty to 100 times as many mobile POPs in a metro as there are wireline POPs, and with LTE each of these will have more data feed than a whole CO did fifteen years ago. Seamless handoff between WiFi and LTE also create changes in topology, and CDN evolution is pushing caching out toward the edge. What we need is a new metro infrastructure where all of the services that we create for mobile and content are created virtually from a mixture of SDN and NFV technologies. I’m talking about “MaaS” or “Metro as a Service”, in effect. If Cisco is doing well in mobile, they need to be talking MaaS to leverage that success.
The data center is another example. Even leaving the cloud aside for the moment, we have this data-center-as-a-service thing going on in the media, which is for the most part vacuous hype and nonsense. Underneath it, though (as is the case with most vacuous hype) is a kernel of truth that could be profound if anyone took the time and effort to deal with it. We need to be thinking about a future where security, access rights, application coupling, big data, governance, business policies, and all that good stuff are attributes of the platform and not glue-on-and-hope-it-works extensions. We need to replace data busses with knowledge busses, replace “big data” with “big insights”. We’ve gotten so used to the idea of managing complexity that we’re risking slipping into creating more of it just so it can be managed. The network is a necessary arbiter of information exchange, an element in every horizontal and vertical flow and every IT-driven productivity process. Cisco is in a strong position to make a name for itself by creating DCaaS that means something.
And then, of course, there’s the cloud. Let me spell things out for you here, Mister Chambers. If you look at all of the things that could drive cloud data center deployment incremental to current deployments the only one that sticks its head above the statistical noise level is Network Functions Virtualization. When we run our model on this, as I’ve said before, we find that an optimum NFV deployment would add between 14 thousand and 35 thousand metro-cloud data centers (small to mid-sized to be sure, but still data centers) in the US alone, about three times that globally. If you want to succeed in the cloud, all you’d need to do is to win big in the optimum NFV space and you’re there.
Cisco has an NFV strategy they’ve been quietly developing in dialogs with Tier One carriers. They need to expand it, to embrace the full potential of function hosting not only for the current services built on current middle-box technology but also for future services where there are no current box models to draw on. The wonder of virtualization as an abstraction/instantiation process is that you can create new abstractions and instantiate them with many of the components you already have. That would provide service agility to operators, and better top-line growth. Top-line growth equates to bottom-line cost tolerance—more spending on infrastructure.
Cisco knows about cost management. They’re cutting 4 thousand jobs to improve their profits in a market where they believe they can’t raise revenues enough to provide the needed boost. Why would they think their operator or enterprise customers are dancing to a different financial drummer? Don’t they also have to face quarterly calls? Won’t they respond to a lack of convincing revenue growth by cutting costs, and isn’t buying Cisco gear a “cost” to those buyers? It’s time to pull your head out of the sand here. Arguing that an Internet of Things or immersive telepresence or whatever is going to drive up traffic inexorably, and force operators to build out even at a loss, is as stupid as arguing that the need for employment drives Cisco to employ regardless of whether they can earn a profit on the result.
Networking is commoditizing, Mister Chambers. Your buyers, in both the carrier and enterprise space, would be happy to spend more to get more return, but they won’t spend more while getting less. It’s your responsibility, Cisco’s responsibility, to make it possible to justify more spending, and to justify it in a solid financially sensible way—the way you justify your own.