If you're looking to incumbent networking vendors for inspiration and leadership about the next generation of networking, it looks like you'll be disappointed. The harsh reality is that none of the big networking companies has a software-defined networking (SDN) strategy. They aren't even sure if SDN is the future or a big pile of stink.
The last quarter has seen a widespread exposure of OpenFlow to the mainstream market, squeals of joy from many market segments and customers starting to seriously consider how it might affect their strategies. The major vendors have started to respond. Brocade announced its OpenFlow plan by announcing that existing products will support OpenFlow APIs. HP has been a major force in OpenFlow, with ProCurve/E-Series switches (not the 3Com/A-Series products) support from the earliest days. Juniper has already announced Junos API extensions and source code to provide access. And while Cisco hasn't announced anything yet, the foundations are there in its XML APIs in Nexus and IOS-XR. Therefore, I'm expecting something at Cisco Live next week as part of usual announcement blitz.
What we aren't seeing are announcements about SDN. The vendors aren't outlining their visions of how we will use OpenFlow. "Oh, we support OpenFlow," they say, but there aren't any suggestions as to how it will be used.
Using OpenFlow means you have to have SDN technology, too. SDN will consist of two broad technologies: controllers and applications. OpenFlow and Netconf will handle the configuration work that the controllers will manage. But the missing piece of the puzzle is the application, which must decide which OpenFlow entries are sent to the network device. The application will analyze the Netconf to comprehend the configuration of the network and then determine forwarding paths through the network. It will decide which path will be used according to some user interaction.
So, where is the innovation? Where is the SDN strategy coming from the industry leaders? I'm not seeing it.
OK. So all the vendors are announcing OpenFlow support. It's cheap to add a bit more software to the existing systems. No pain, let's tick that off. And given the number of articles and wide-ranging interest in OpenFlow, it would be relatively simple to convince the marketing and product managers to get this done quickly.
On the other hand, setting out to develop and announce an SDN strategy? Whoa! That's gonna take some time. Big vendors will have to check with customers for "use cases" and determine if there is "market demand," "evaluate the internal strategy," and then make sure there's "program buy-in" with "management engagement." Instead of innovation, we're getting bureaucracy, resistance to change, and dull, boring marketing routines repeating the same old story.
Vendors should grab hold of this opportunity to deliver leadership and inspiration to customers. Look at NEC. A year ago, NEC didn't even exist as a networking vendor outside of Japan. Today, it has partnership deals with IBM and Brocade, and is positioned to claim leadership of an entirely new networking market. Compare that with the Cisco and Brocade announcements about their SDN controllers ...
Once upon a time, the big networking vendors would develop exciting products and features. Then they would preach them to their faithful customers. This is how many technologies got into the market. Remember technologies like Cisco's ISL, which drove the use of virtual LANs. Did Cisco ask customers, "Do you want this?" and spend two years developing a product plan? No, it went and did what was obviously needed.
VMware created innovation by changing the server market five years ago. SSD startups are changing the face of storage. Apple is changing the user experience to a Post-PC environment.
And in networking, here's the message from the vendors: "We aren't convinced that SDN/OpenFlow is something real yet. We need to wait and see if it's a real thing."
Ten years of stagnant technology in networking, and a chance at a revolution! I could puke with frustration.Greg has nearly 30 years of experience as an IT infrastructure engineer and has been focused on data networking for about 20, including 12 years as Cisco CCIE. He has worked in Asia and Europe as a network engineer and architect for a wide range of large and small firms in ... View Full Bio