6:05 PM -- Software-defined networking (SDN) stands out as the singular concept that is both hard to define and hard to stop talking about. Dozens of companies that provide enterprise and cloud infrastructure want to separate the software elements that control the network from the network hardware itself.
The big idea -- and this is a classic tech industry argument -- is that you can accomplish more by separating the operating system from the applications, the operating system from the hardware and, in cloud computing, the server from the service. Based on the conversations I've had this week at Interop, my feeling is that SDN -- of which the Open Networking Foundation's OpenFlow is one example -- is nothing more than an industry uniting against entrenched powers and looking to disrupt the hold of one or two vendors that are so dominant that network innovation seems to move at the speed of their product cycles.
Would OpenFlow have momentum if cloud service providers felt like the top networking infrastructure vendors were listening to their customers and building switches and routers in a way that gave them control and the flexibility to create and provide new services?
Is it surprising that Amazon.com Inc., Google and some the earliest examples of major public cloud providers all wrote their own software and network stacks and used non-traditional infrastructure to start providing network services to their customers? Why? Couldn't they find a vendor to give them a cloud solution that was priced with cloud economics in mind?
This isn't to say that the networking infrastructure providers aren't cloud-capable. We ran a monster test of various Cisco Systems Inc. cloud capabilities and you can see for yourself that Cisco does have a cloud(y) story. Cisco wants to give service providers the tools to handle delivery of any kind of content to any kind of device, where and when subscribers demand it.
But do service providers always want all those tools to be Cisco tools, delivered the way Cisco wants to build them, on Cisco's product development schedules, with Cisco's software controlling every aspect of every service? A lot will, actually. But what if some of those providers want to know if there's another way?
Each vendor I spoke to at Interop had its own take on SDN. John Roese, senior VP and general manager of Huawei Technologies Co. Ltd.'s U.S. R&D Center & Enterprise Business, told me not to get too excited about OpenFlow. It's a feature, not a product, he said. Oscar Rodriquez, Extreme Networks Inc.'s CEO, told me that his company is better positioned to adjust the flexibility of a software-defined network because all of its products run a single operating system, a contrast to companies that grow by acquiring dozens of products and trying to make them all interoperate.
I doubt SDN will kill today's equipment giants, but it might make them more flexible. And it is safe to say that the value in this industry is shifting away from hardware. The virtualization of assets, and the control of that process, is just one element getting more attention. You can add up and double the value of Infinera Corp., Riverbed Technology Inc., Alcatel-Lucent, Extreme Networks and Juniper Networks Inc., and you'd still need $6 billion to match VMware Inc.'s market capitalization. While we're measuring market caps, Cisco, with seven times as many employees, is only twice as valuable as VMWare.
So, yes, the buzz around SDN may be hard to sum up. But can it be measured? Big Switch Networks's OpenFlowHub, a site that hosts open-source networking projects based on OpenFlow, is getting more than 28,000 page views per month. If today's networking innovation was moving as fast as the clouds forming overhead, I don't think that'd be happening.
For more
— Phil Harvey, Editor-in-Chief, Light Reading
