Twenty years ago, enterprise networking as we know it was born. IBM, whose System Network Architecture (SNA) had dominated networking from its birth, lost its grip on the space. SNA was swept away, IBM got out of the networking business, and Cisco and other network equipment vendors rose to prominence. The networking space has had a great run, but whether that run continues will probably depend on the outcome of the latest war in the data center.Money is what transforms enterprise technology-spending that's justified by benefits that information technology and networking bring to business operations. Productivity tools are a combination of information, processing, and networks that connect this mix to workers. When computing advanced faster than networks, as it did through the '80s with the PC revolution, networking had to explode to catch up. But applications really support productivity, and now there are some indications that connectivity isn't the limiting factor in product benefits any more. Instead, the limiting factor may lie in the data center, and that's changed the competitive dynamic of the industry.
Remember how we used to have "software vendors" and "hardware vendors"? Now we find IBM and HP getting more into software. We find Oracle getting into hardware. This broadening of company focus is happening because the projects of the future--the ones that will define IT spending--are projects that require changes in both software and systems.
Enterprises aren't completely confident in the new paradigms, which include things like SOA that's been around for half-a-decade and cloud computing and virtualization that are pretty new. A lack of confidence in the buyer community translates to a more educational and difficult sell. Why should a vendor who spends sales cycles to convince a buyer to revamp the hardware, middleware, and software of their data center then settle for selling only one of those things? That's what's fueling the spreading out of the big players.
Networking is a part of the data center too. In fact, networking may be the critical issue for companies looking to get into virtualization on a large scale, and into cloud computing on any scale. But the networking issues of virtualization and the cloud are linked tightly to the other components of the data center, and so networking is being drawn into the big picture. That's why HP is buying 3Com, why IBM did OEM deals with Brocade and Juniper, and why Cisco got into the server business. The data center isn't technology pieces any more, it's one glorious technology whole. And that glorious whole will go to a kingpin vendor who can put it all together in the first place.
We could easily end up with a future where networking is subsumed in IT, where connectivity in terms of switches and routers isn't any more differentiable among vendors than connectivity in terms of cables and jacks and plugs. Whether we do or not depends on what two vendors do--Cisco and Oracle.
Cisco is the network guy turned server player. Most on Wall Street believe Cisco read the writing on the wall and dived into servers to increase their total addressable market (TAM) and thus allow them to grow faster than the networking industry. But if Cisco lets networking as a market commoditize faster than it can pick up server market share, it loses revenue and its street (Wall Street, that is) cred. So Cisco needs to make things like virtualization and cloud computing into networking issues and not IT issues.
Oracle is on the opposite side. They're a software giant turned general-purpose IT giant with their Sun acquisition. Their move into hardware could force Microsoft to think about getting into hardware itself. Oracle's move into the cloud and the Internet in an application and server sense (Sun is a giant in both the Internet and in telco IT) could encourage them to get more into the networking business too. There are rumors that Oracle is looking to buy a network equipment player. What would that do to the industry?
The only player in the New Data Center race that wants networking to be strong is Cisco. IBM and HP see it as an adjunct to IT and Oracle and Microsoft aren't committed. If Oracle gets into the networking business it joins IBM and HP in their networking-is-secondary position, and that's enough to drag Microsoft along. Every one of these multi-purpose giants would then oppose Cisco in making networking a differentiable element of a future IT plan. It becomes connectivity plumbing.
The power to stop networking from sliding into an also-ran role in the data center and the IT budget lies with three vendors; Brocade, Cisco, and Juniper. These are the three that have specific data center networking (and in Cisco's case, data center/server) products that could create a compelling vision of how networks create clouds and support virtualization. No, that doesn't mean the tired old stories of "connectivity" and "QoS" either, it means creating an application layer, a form of "cloudware" that could revitalize the role of the network and the people who buy and support network products in the enterprise.
IT progress doesn't come via a rank of technology moving ahead in unison, it's a bit of a disorderly race. Right now, networking is starting to run behind the pack, and the users and vendors who depend on networking need to take some steps to establish the network as the framework of the "virtual data center" of the future. Otherwise the data center is going to become the virtual network.
