Interop continues to dominate the news this week especially around the SDN buzz generated at the conference. Many remain divided on whether companies will make the necessary changes for network improvements, particularly within their IT organizations. We think operational necessity will drive the transition and the SDN market will reach $35 billion by 2018. Plexxi is excited to partner with Calient Technologies and SDNCentral for a DemoFriday on June 14 to discuss SDN-driven, dynamic packet-optical datacenter architecture. Please register if you haven’t already! Here are a few of my reads for this week in the Plexxi Pulse – enjoy!
Tom Nolle from CIMI Corporation reviews Juniper’s Contrail details from Interop and the “JunosV Contral Controller.” Tom has been searching for what might distinguish Juniper’s approach from others. Most of what transpired at Interop was talking about the “why” followed by a bit of “what” (and a protocol spec or two). Today the industry is still continuously trying to confuse the minds of customers: It does so by quickly switching back and forth from “what” to “how” without touching any of the “why.” Unless you are a protocol designer, having an “XMPP is better than OpenFlow” conversation is largely irrelevant. The mechanics (“how”) are much less important than what you do with it and why. Once we agree on what information needs to pass, and at what abstraction level, then the industry can move quickly to technology. General hygiene requirements will weed out technology that does not fit a customer’s requirements. IPSphere’s problems were not technology at all – technology is an excuse to steer clear from the hardest topic of all: Information monopolies. What information is mine, what is yours, and what is it we should share to get our joint jobs done? Side-stepping that conversation by talking protocols keeps the industry focused on the wrong outcome.
John Dix from Network World says SDN was the hot topic at Interop earlier this month. Conference attendees got to view a parody movie trailer for the arrival of SDN, and while there was widespread optimism, John notes that the Interop audience admitted to being less prepared for a major implication of SDN–change within IT organizations. It’s great to see the SDN momentum continue, but as John suggests, I wonder how many companies are looking at things like organization design, budget process, decision-making process and talent as they consider SDN. I don’t believe the impacts are purely technological and those companies that get ahead of these aspects will likely make more significant progress. If SDN is indeed strategic for companies, it might be of interest to look at the change management elements necessary to pull this off. I fear that organizational impacts might blunt the positive impact at best and dissuade some folks entirely at worst.
Enterprise Networking Planet’s Arthur Cole believes that while it seems there is little preventing the emergence of virtualized data centers that can be commissioned, utilized for either specific or general purposes, and then decommissioned, entirely in software, the enterprise isn’t truly ready for SDN and the hype it has generated. SDN is still in its infancy according to Arthur. I’m curious where Arthur stands on the talent implications of this highly orchestratable environment. Do companies have the expertise on hand already to take advantage of SDN? Will the role of the network engineer evolve? Or will new people come in to do new things? Some questions to think about as we engage SDN.
Kevin Fogarty from Next Generation Data Center also discusses the SDN energy from Interop in Las Vegas. Kevin thinks SDN’s promise to upend traditional networking to the same degree of virtualization transformed servers and applications drives the excitement. One of the measures of SDN being hot is just how many companies self-associate with it. I think SDNCentral reported something like 225+ companies identify with SDN, up from 0 in 2009. And if VC money is any indication, both enterprise and networking are hot again. LightSpeed reported a 50-fold increase in VC money moving to SDN companies in 3 years. One of the interesting impacts, in my opinion, is that SDN will attract a new flock of talent that had previously been uninterested in our space. Money brings in the youth, and we need new people with new ideas to avoid incrementally moving the networking forward as we have done for the last 10 years. Great stuff!
Tom Nolle contributed to No Jitter this week to discuss which SDN was the real SDN. Whether it’s about OpenFlow, MPLS and BGP or rich APIs, vendors see their own SDN as the winning SDN. VMware, for example thinks it’s about software virtual networking, but Alcatel-Lucent thinks that’s only part of the story. “There’s a deep SDN and a shallower one,” says Tom, and I like the approach of deep versus shallow. Some might say that shallow is really what network virtualization (in the overlay world) aimed to solve. In parallel to questions about the utility and benefits, there is a discussion we have to have about value: low cost does not mean zero cost. If the perceived value of “shallow SDN” is low, then having this be low cost should not be a problem. However if you believe that managing this problem space is of high value, then the subsequent solutions to the market might not be “low cost.” Comparatively speaking it might be “lower cost” than sheet metal and silicon, but for some definition of value there is going to be a price tag associated with it, even if that price tag might be under pressure.
The post Plexxi Pulse – Looking ahead to the June 14 DemoFriday with SDNCentral and Calient Technologies! appeared first on Plexxi.
