It should be clear to everyone in networking this week that the industry is moving to a very different place. There was a time when the limiting factor in entertainment, communications, and empowerment was delivery. We spent like sailors (as they say) to get better delivery because we had a business case for what we could deliver and it was hampered by the limits of network capacity and connectivity. Practically nobody bothered with a strict business case. I remember when financial analysts valued an Internet customer at $50 thousand, and valued Internet companies by multiplying that by the subscriber base. The good old days.
This morning on one of the financial channels, a tech markets specialist was talking about an interview with Cisco’s CEO, and he commented that he didn’t understand how Cisco harmonized the fact that the industry was supposed to be getting better according to Chambers, but oh by the way we’re cutting staff. I think the answer is pretty clear; Cisco knows darn straight well that the industry’s not going to sustain their growth expectations. Business cases have caught up with networking…with tech, in fact.
We are starting to see in networking what we’re seeing in computers, which is an exhaustion of the value proposition for further investment. Clearly we’re not going to toss servers out and go back to high stools and accountants wearing green eye shades pouring over ledgers, but we are already looking for cheaper servers just as Chambers is looking for lower human costs at Cisco. ROI is the ratio between return and investment, and if you can’t eke out any more return you’ve got only one way to make ROI better.
In the computer space, we heard this week that Dell had gained on both IBM and HP, increasing its server sales when rivals saw theirs dropping. I think that underneath all the disorder created by potential going-private or getting raided, Dell is facing the transformation in computing as well as anyone could. In fact, in some ways, the current mess sets Dell up with some ashes from which its Phoenix can rise. All they need to do is to face the future squarely.
Which is what? In my terms, it’s the “supercloud”, an architecture for computing, networking, service creation, service delivery, that focuses on agility through virtualization. Right now, we as an industry are stuck in conventionalism. It’s like giving a kid a chance to play doctor, or lawyer, or police officer or maybe soldier, and they decide to play “wage slave” instead. What good is imagination if all you can do with it is replicate the current reality? Cloud computing, SDN, NFV are all simply applications of virtualization, the realization of flexibility through the creation of a dynamic system of “abstraction-creation” and a generalized way of converting abstractions into reality by making dynamic resource assignments. That’s what we should be doing, and we’re not.
The cloud lowers costs, they say. SDN lowers costs, they say. NFV will lower costs, they say. All this lowering is just cutting the heart out of the industry, and for no reason. We have, by my reckoning, almost three trillion dollars per year of available benefits to harness. Dell, I think, sees that pie (or a piece of it) and is thinking about how to get it. Who, in networking, can say the same?
Ironically, the vendors that seem to have the best grasp of the current reality are the vendors who for one reason or another would seem the least likely to have any grasp at all. Alcatel-Lucent is a company that’s struggled with multiple personality disorder from its formation, and that’s never been able to tell an exciting story except maybe over a drink. Huawei is a company who’s going to win in the current networking game because all its competitors are too dull to recognize that if there is no revolution in networking’s future, the future will be price-based commoditization that only Huawei can hope to win. That Huawei is already winning. Yet Alcatel-Lucent has taken some concrete steps toward the supercloud model with CloudBand, and Huawei is really looking at software-defining the network.
Network operators believe in the supercloud concept even if they’d not likely call it that. They believe that by moving away from static appliances into virtual elements they can improve agility and service creation, and thus address opportunities faster and better. They believe that the OTTs have nothing on them except the ability to exploit “free” services, and ad sponsorship is both not-free and inherently limited. The whole global advertising budget is about one quarter of that available revenue upside for the industry. Operators know how to sell services; OTTs know how to give them away. In the end, which model wins?
I think the industry could have an exciting future, but as long as enterprises have the CIO reporting to the CFO instead of the CEO and as long as “transformation” means radical cost-cutting, we’re not going to reach it. And this is the fault of the vendors, because there has never been a free market that succeeded by demanding that the buyer go out and evangelize to convince the seller to produce something. Innovation, revolution, doesn’t work by demanding buyer risks. Watch “Jobs” if you don’t agree with that.
People tell me that SDN isn’t going to work, and I agree. Same for NFV, and for the cloud. But my agreement isn’t based on a conviction that these concepts can’t work, but on the conviction that the goal we’ve set for them—the goal of wringing some pathetic chump change from re-execution of the tired junk we’ve relied on for decades—can never inspire the market. We have tied virtualization, as a concept, to a tree and then criticized it for not getting around. Well, we’re the problem here, and so we as an industry have got to fix it by fixing ourselves.
