Question: Will Vendors Open Up About OpenFlow at Interop?

It’s the week before Interop so you are likely to face a fortnight of public relations buzz from the networking industry. We’ll hear plenty about FCoE and its supporting parts (Converged Enhanced Ethernet (CEE), Data Center Ethernet (DCE), there should be lots of hoopla about “flat networking” standards (Trill, SPB) as well as proprietary alternatives, and we’ll likely face a barrage of new product announcements as well.I have lots of meetings set up to discuss these industry happenings, but I plan on asking about another potential networking game-changer subject with every company I meet with. I want to know what each networking vendors thinks of OpenFlow, whether it plans to support OpenFlow, and in what time frame. Just what is OpenFlow? It’s a lot of things but for my purposes, it is a set of technology standards that externalize the control plane from a standard packet-forwarding network (note: My friends at the Internet Research Group have published an excellent synopsis of OpenFlow and its impact on the market). OpenFlow can still be classified as more science project than real-world implementation but it has the potential to really change how networks are built, managed, and sold. Some vendors stand to gain by these changes, some see OpenFlow as a potentially huge threat to their businesses, some aren’t sure what impact it will have, and some choose to ignore OpenFlow all together. That’s what makes it so intriguing. I’m sure that Network World’s Jim Duffy will be following OpenFlow as part of his Interop coverage and I plan to do the same. Here are a few of my thoughts before the show:1. I haven’t had the chance to ask Cisco about OpenFlow but I’m most interested to the answer to this question. From what I’ve read, Cisco joined the Open Networking Foundation (ONF) that owns OpenFlow and is saying publicly that it is cool with OpenFlow as long as vendor extensibility is part of the specification. What does that mean? Stay tuned, that’s exactly what I plan to ask.2. I’m meeting with Citrix and VMware next week, I’m sure that each will be big proponents of OpenFlow. My guess is that Microsoft is too but I find it nearly impossible to find the Microsoft networking people these days so I won’t be meeting with anyone in Vegas. I wonder if Microsoft is planning to support OpenFlow in either Hyper-V, Azure, or both? I’m sure that other software vendors like Oracle, SAP, and industry providers like Meditech also see benefits with OpenFlow. 3. Juniper has obviously gone in a different direction. QFabric mimics an OpenFlow architecture but does so in a proprietary way. Juniper is also a member of ONF but given the investment in QFabric, I can’t imagine that it will take an active role. Nevertheless, it is worth asking.4. Aside from Trill and SPB, wouldn’t OpenFlow accelerate the transition to flat networks? I have a bunch more prepared but please send me additional questions to fire at the networking industry.OpenFlow has the potential to really change the networking industry just as server virtualization is changing application, server, and data center deployment. That said, it is also important to keep things in perspective. Many users are struggling to keep up with the massive changes brought on by server virtualization. They often lack the right skills, best practices, and organization to take advantage of the dynamic nature of server virtualization and cloud computing. The same holds true for OpenFlow. Yes, it could change the dynamics of networking but most users want switches and routers that just work. If OpenFlow can get to this level of pragmatic functionality, simplify network architectures, and streamline network operations it may be a “sure thing.” If OpenFlow gets “embraced and extended,” demands new skills, and required complex network programming, it reside in University labs and cloud computing alone.