Cisco reported roughly in-line numbers for the quarter but the stock was down over 2% because it is still reporting a sequential decline in revenues. Guidance was also at the low end of Street expectations, which further suggests a “no-improvement” scenario, and there was special Street concern for the fact that security products didn’t do as well as expected. I’d guess you’d not be surprised if I said this wasn’t a good sign for networking.
Why does equipment revenue decline? Because operator and enterprise capex is at least not growing, and overall is declining. Why is that happening? Because it takes new benefits to justify new spending, and buyers are focused on reducing their own costs. Cisco and other vendors are cutting costs so their profits aren’t dropping with revenues. Buyers are supposed to do something different? Get real.
If you’re not surprised that I don’t think this is good for the industry, you won’t be surprised if I say that it’s hardly news. Operators have told the Street that they’re cutting capex. Enterprises don’t have that kind of long-term capital-project planning, but CIOs are telling me that every round of new network spending is focused on lowering costs overall, and if that can’t be done then at least raising them as little as possible. All of this because ROI for network projects isn’t meeting internal guidelines for approval unless the cost is lower, not higher. Lower cost, lower spending, lower vendor revenue. QED (which for those not blessed as I with two years of high-school Latin, means “quod erat demonstrandum”, or in English “Thus it has been demonstrated.” Or, if you like “res ipsa loquitur” or “the thing speaks for itself”. Don’t you love a practical education?)
I really feel like “Groundhog Day” here, as long as we’re doing quotes. Yes, I have been saying that absent new benefits there cannot be new spending. No, people don’t seem to be paying attention. So I’m saying it again now and perhaps offering some comments based on Cisco’s earnings call comments.
Vendors love to explain shortfalls as being due to market conditions and not to their own missteps. Well, gosh, is it a surprise to them that they are in the market? What conditions did they expect, and did they have anything other than blind hope that they’d come along? If you dig past the usual trash in vendor comments, they all are saying in effect that they thought that more traffic would drive more earnings in the operator space, and also in the enterprise space. More bits means more bucks, and that’s true for vendors. Not so for buyers.
The future is never a linear extension of the past. Any technology, any business idea, has a logical lifespan beyond which its benefits no longer grow, and so no longer justify increased investment. We are today depending on a notion of networking that goes back about 40 years. What in tech has survived that long? In 1974, the year TCP/IP arguably was born, a small computer was one that would fit in a 19-inch rack and cost thousands of dollars. Today a computer a hundred times more powerful can be carried in your pocket. So how could computing change so much and networking change so little? It’s not logical.
To be fair, though, computing’s change was more quantitative than qualitative in networking terms. Today’s systems are a lot faster, but they are still discrete devices that have fairly static relationships with networks. A network that connects a multi-core smartphone and a network that connected a DEC PDP-8 still address endpoints the same, and expect connectivity between those endpoints to be the essence of any service. Cisco and others, perhaps, might be forgiven if they fall today into the same service-mission trap that ensnared them four decades ago.
Can ignoring four years of static or declining revenue be forgiven, though? Certainly it shouldn’t be ignored. I saw the classic “profit-per-bit” compression and crossover slides in 2013, and so did a lot of other people. Cisco now, perhaps more than other network equipment vendors, is ready to face the truth and push more for “software and subscriptions” as a revenue source. The question is whether this shift can really accomplish what Cisco needs.
All my modeling and all the logic in the industry suggests that networking and computing had a kind of push-pull relationship initially. Computing creates an information/content pool that, for a time, was inhibited by network infrastructure designed for low-speed voice and terminal traffic. Cisco took advantage of the sudden excess of that-to-be-delivered and insufficiency of delivery options. Now the problem is that we need more from the compute side—processing resources to enhance the value of delivery again. And it’s not obvious how that happens.
Corporate IT needs to reframe its network to support point-of-activity worker empowerment, creating what is effectively an event relationship with workers. Consumer services need to be able to contextualize every consumer interaction to make them more valuable. I know that Cisco knew there were at least some who said this a decade ago, because I told them. They, like most of the industry, elected to stay the course of traditional networking.
Security is another example of short-thinking. Do we really think that network operators and enterprises will pay nearly as much to secure networks as they paid to build them? Is the fact that, as Cisco says on their earnings call, IoT device security attacks are up 90% an indication that we need to spend gazillions of dollars on IoT security? We need to be spending more on making networking intrinsically secure, not gluing remedy onto imperfection.
Cisco’s call says in essence that data center is growing and WAN is flat, and they correctly name the cloud as the reason. However, where in the call does Cisco say they know why the cloud is growing in importance? It’s not because it’s cheaper for current applications, but that it’s the right platform for future applications. There’s a lot of computing changes between us and where we need to be, in order to support those future applications at the server/software level. That’s where Cisco, and other network equipment vendors, need to be. Don’t expect the consequences of cloud expansion to win the game for you, expect to win it by driving that expansion directly.