Cisco is in the fortunate position of the rest of the world having moved to where it already was.
Most companies looking to plot a long-term strategy attempt to figure out where the world is moving to and then position themselves slightly ahead, ready to meet customers with products and services they want. It’s a great theory, but notoriously difficult in practice. More often, the world changes and companies scramble to keep pace with it. A few years ago it became accepted fact that everything was going to public cloud, software was where the money was, not hardware, and everyone had to become a software company or die.
Everyone was saying so, so it had to be true.
Cisco, responded as many did at the time: it tried hard to convince everyone that it totally was a software company because that’s what investors and VCs were pressuring everyone to be. Software margins were very attractive, and no one wanted to be in the uncool business of something as prosaic as hardware. So Cisco, like many other companies, duly trotted out messaging designed to respond to this fashion trend while not really delivering very much in the way of actual product or business changes.
And then fashions changed.
The world decided that, actually, no, putting everything in one cloud isn’t what we want, we want lots of clouds, and also datacenters aren’t going away after all, oh and colo facilities are still useful and also stuff out at the edge is going to stay there and in fact grow quite a lot.
Happily for Cisco, all it needed to do was to change the way it talked about what it was already doing, and voilà!, Cisco had clearly planned it all along.
The tech industry has an unfortunate preoccupation with the new and shiny and its infatuation with software put Cisco in a difficult position. “We’re unique,” Suzuki said, “Because we’re never going to give up on hardware. We’re never going to be one or the other.” Five years ago this statement would have been decried as Cisco being outdated and out of touch with the Way Things Are Done Now. Today it’s a fairly innocuous statement that accurately reflects what Cisco does. Software needs to run on hardware, and certain software workloads benefit from specialized hardware. That’s why clouds, for all they bang on about being about software, have been loading up their virtual machines with GPU cards to do machine learning.
Hardware matters. And yes, so does software, because that’s what runs on the hardware, so having a bit of both is actually quite handy. Modern hardware is actually full of software anyway, it’s just embedded into the chips where you can’t readily see it so everyone sort of pretends it doesn’t exist most of the time.
“It’s not like we’ve never had software capability,” Suzuki adds, “It’s more the business model is becoming more software-like.” Mostly this is about financing rather than functionality. “Increasingly we’re finding customers want things on a recurring subscription model and also as-a-Service so that they don’t actually have to manage the asset,” Suzuki said. Customers will still have the same physical devices and software from Cisco as they would have if they’d purchased it outright, they just lease it instead.
And there’s nothing wrong with that.
Cisco is fairly well positioned for an explosion in edge computing devices that need networks made of hardware and software to tie them into the rest of the global computing environment. Whether this is due to luck or good management is an open question. History teaches us that luck is probably a larger component than we feel comfortable acknowledging. Still, Cisco has accumulated plenty of expertise in a variety of product areas that are important in the multi-cloud/hybrid/edge computing world we find ourselves in. The vast amount of networking required to connect all this distributed computing stuff together represents a massive opportunity for a company that certainly knows plenty about networking.
Cisco has also been steadily adding standardized APIs to its products so that developers (and network admins) can learn how to program them and manage them using more modern, automated approaches. Certainly this could have been done faster, but it seems that networking folk are somewhat slower to adopt new ways of working than other parts of the stack, which surprises me but appears to be true nonetheless.
Still, I can see enterprises using these interfaces in combination with tools like ServiceNow to add more self-service and automation in their business processes now that automation has become sufficiently fashionable. The prospect of networking infrastructure that is programmable without having to purchase additional vendorware is attractive, particularly if the vendor’s tools don’t work the way your business needs them to. If being API programmable is a Cisco competitor’s chief advantage, that advantage will be gone shortly, or at least severely blunted.
Cisco has been quite lucky that the majority of IT customers moved about as quickly as Cisco did in swapping everything to the cloud.
The challenge now for Cisco is to capitalize on this lucky break. Most organizations don’t even get one of these, let alone two, so to squander the golden opportunity before it would be a tremendous waste.
I attended Cisco Live in Melbourne as a guest of Cisco. You can read my full disclosure here.
This article first appeared in Forbes.com here.