Why it makes sense for Broadcom to buy VMware

Why the heck would a hardware and chip company like Broadcom buy a software company like VMware?

Wall Street and industry analysts haven’t exactly jumped with joy over the pending deal, after all. Companies sometimes do stupid things; that seems to be the consensus. But with this deal, that may not be the case at all.  Broadcom may be responding to the fundamental shifts in the industry, both in computing and in networking.

The cloud has done a lot of things to computing, but one of the most important has been overlooked. It’s given applications identity. We now think of application front-ends that customize customer, partner, and worker information. We think of application hosting, application orchestration, application visibility and performance. The data center used to be the focus, the resource that ran everything. Today, I’d argue that virtualization and the cloud have gotten us back to where our focus needed to be—the applications that support businesses.

The cloud has also magnified a trend in networking that started decades ago with VPNs. Companies don’t build wide-area networks these days, they connect to them. VPNs eliminated the buy-your-router-nodes-and-connect-leased-lines approach, and the natural integration between the cloud and the internet is moving us even further (and faster) away from WAN equipment. You access a service; you don’t build it.

Hybrid cloud: The application front end

But the cloud isn’t all of computing, and that’s proved by IBM. They’ve bucked the recent tech route, and one big reason for that is their relentless focus on hybrid cloud, the application-hosting model that uses the cloud as the front-end user interface to legacy applications that still run in the data center and work with critical corporate data that’s also stored in the data center. Hybrid cloud requires access to the cloud—a network function—but also connections within the data center. It’s a driver of change in the data-center network.

Enterprises do build their own data-center networks. In fact, the data center network is the critical link in hybrid cloud because it’s what links the front-end and back-end pieces of applications, the most mission-critical applications a business has. The connection between cloud and data center is the pathway that all critical data follows, the portal through which hybrid-cloud orchestration is controlled. All of this passes through the data-center network.

The network trend, the shift away from building a WAN, means that the data-center network is probably the only real network an enterprise will build. It’s the place where capital equipment is purchased, where operations is totally the responsibility of the enterprise. It’s also the place where the hybrid cloud is anchored, in (you guessed it!) applications.  The data center and its network are the linchpin. Broadcom’s switching chips are already dominating white-box data-center and SDN switching, and now they want the application connection, meaning VMware.

By combining the chips that are the foundation of a data-center network and the software that creates the application platform, Broadcom has a solid position in one of the two technology points where the hybrid cloud lives. The other point, of course, is the cloud.

Let’s play “suppose” here. Suppose that the cloud is the future of IT, but suppose that IBM is right in its view that hybrid cloud is the flavor of cloud that matters. That means that the data center and its platforms and applications are more important than we’ve thought. Maybe they’re more important than the “cloud” piece of “hybrid cloud”.  All the publicity, all the media focus, all the developer interest, has been flowing to that cloud piece, leaving the data center sitting like a nerd at a dance. But they say that nature abhors a vacuum, and all that interest and attention departing the data center seems to have created one. With everyone focused elsewhere, might this be a good time to frame a powerful data-center strategy? Suppose it is, and suppose that’s what Broadcom is doing.

Almost all mission-critical applications and data are still in the data center.  Enterprise CXOs tell me that’s not going to change any time soon, if it ever does. They say that the cloud-centric mindset these days comes from a combination of media focus, no new stuff, and the fact that the focus of new development and new technology planning is on the part of hybrid cloud that’s also new, which is the cloud part. The important stuff is what the cloud hybridizes with.

The cloud will at some point become commonplace. The media will move on to quantum computing, neutrino computing, or something equally esoteric. The applications and the data that anchor hybrid cloud will remain, so the data center will remain as well.

Continuing with our suppose game, suppose that this is what IBM figured out. Suppose the cloud providers have figured it out, too. Suppose that a common platform technology for applications in both cloud and data center is going to become critical, which is why they’re all working hard to extend their platform technology into the data center. VMware has the platform tools to counter that. Suppose that data-center networking is the other critical piece. Broadcom chips could be the key to that. With a single acquisition, Broadcom closes the loop.

Computing in data centers goes back 70 years at this point. We’ve had transformation after transformation in the data center, we’ve had computing shrink down to something the size of a watch, we’ve had a shift from real to virtual resources, and yet we still have data centers. Remember when distributed computing or personal computing were supposed to eliminate them? They’re still here, and still critical, and the data-center network is both here and critical, too. A VMware acquisition won’t assure Broadcom success in this area, but it will tee up the possibility.

When I was a boy, I read a story about a lad who was criticized for fishing in what was the local junk pond.  When confronted, he told a long and complicated story justifying his decision, which boiled down to the point that it was possible that the pond was a gateway to a vast fishing opportunity. The story illustrates that you can stretch possibility too far. The story ends before the issue of the fish is resolved. Maybe Broadcom is seeing possibilities and not real opportunities, just like that boy, but just suppose. Suppose they’re right.