Leveraging the fact that there were so many SND companies, principals and experts in the Bay Area this week, CCICE sponsored a Software Defined Networking panel discussion last night at the TechLAB Innovation Center in Santa Clara.
The personalities on the panel made for an interesting set of perspectives, as participants ranged from start-up principals, to incumbent networking providers (Cisco), to venture capitalists, to customers that need SND solutions now. While the size of the panel and the complexity of each panelist’s story left little time for questions, the diversity of views presented was great. A brief summary of the presentations follows.
John Vincenzo of Embrane presented a very fast paced overview of Embrane’s abilities to quickly deploy network overlay services such as VPN, Firewall, SSL offload, or load balancing to create added compute capability, simplification and agility to enable better IT management of private or public cloud services. Per John, customers are using Embrane’s services to provide non-disruptive compute elasticity, HA / self-healing, and automation for managing complex networks now.
James Liao, CEO of Pica8, explained his company’s implementation of an open, hardware independent, Network Operating System, PicOSTM, that supports layer 2 and layer 3 protocols and also features SDN OpenFlow support. With Pica8 software running on physical switches, customers can integrate the data plane and the SDN controller layer now – using OpenFlow – delivering significant SDN functionality. A look at Pica8’s website is well worthwhile to get a flavor of their offerings.
Jason Matlaf of Big Switch Networks started off by emphasizing one of the most driving applications for SDN – that of simplifying network service deployment and configuration in cloud compute virtualization. This was generally understood, but not mentioned so far in the discussion. Because compute virtualization is more and more dynamic and flexible, IT personnel are constantly reconfiguring and redeploying network resources to the point that networking is “in the way” and VM resources can be highly underutilized.
Jason went on to give an overview of Big Switch Networks solutions and their two primary applications – Big Virtual Switch (the network visualization application) and Big TAP (the network monitoring application). He emphasized that with Big Switch solutions, cloud data center operators can move from static network segmentation and configuration to dynamic, rules based, and automatically configured network segmentation which allows for much higher virtual machine configuration density (cap-ex savings), dramatic improvements in agility, and the OpEx savings associated with automated and dynamic network configuration. One point I personally applaud regarding Big Switch Networks is their contribution to the open source for SDN, notably with the Open Source core of their Big Network Controller, FloodLight.
Next, Joerg Nalik, IT Director at SAP, provided a customer’s perspective. He made it clear that SDN solves real problems for customers and is not just technology for technology’s sake. Joerg is looking to SDN to help him with – 1) application availability – to minimize downtime and assure customer satisfaction; 2) application security, to assure customer data isolation and protection; 3) to maintain optimal application performance. This, Joerg said would be enough, but he also anticipates significant cost savings and op-ex gains from SDN in the long run. Joerg emphasized that for large data center operators like SAP, the complexity of managing traditional networks is becoming more and more burdensome. With the advances in virtualization in computing, it is very important that similar advances are made in networking.
Arif Janmohamed from Lightspeed Capital then provided some very interesting figures as to why VC’s are so interested in this space. Arif suggested that given the disruptive nature of SDN, there may be as much as $35 billion of revenue up for grabs depending on the winners and losers in the evolution of SDN. This suggests that there could be as much as $150M in valuation up for grabs. Clearly big numbers. He advised us all that given the stakes here, the audience should “think big”. With the massive changes upon us due to the availability of x86 compute capability, virtualization, and cloud computing, there will be many different SDN solutions that will be successful as this new ecosphere matures. He advised the entrepreneurs to look for every opportunity to virtualize “boxes” and push them to the cloud.
Finally Shashi Kiran, Sr. Director of Data Center and Cloud Networking at Cisco, spoke on Cisco’s approach to SDN. Needless to say, with Cisco’s resources, Cisco has a broad approach to SDN. First, Shashi emphasized that while SDN is very important to Cisco, and clearly a technology of critical importance, Cisco has an obligation to support its customers with the architectures and services that they have already invested in, and that continuity remains important. And, Shashi emphasized the services that Cisco is deploying already to simplify configuration and management of complex networks, such as their “onePK” program (http://www.cisco.com/en/US/prod/iosswrel/onepk.html). Shashi also discussed Cisco’s participation in the OpenDaylight work with the Linux Foundation, and the plans for investment and contribution to that effort. But Shashi was careful to emphasize that while Cisco was contributing to the open SDN activities and believed this to be very important, Cisco did mean to maintain a focus on delivering integrated solutions – that will be somewhat proprietary by nature – in order to deliver the best solutions to their customers – while also providing appropriate interworking with SDN standards and capabilities.
Following the panelist’s comments – which took most of the evening – some great questions from the audience did stir up quite a bit of passion from the speakers regarding who had the best solutions, how the start-ups would fare as the ecosystem developed and what would happen as the large OEM players became more involved in SDN. The passion was visceral and exciting to be a part of. But in the end, the panelists for the most part returned to a focal point – that what SDN needs to focus on first and foremost is to simplify the rapid deployment and configuration of networking to support the new demands coming from cloud computing, dense data farms and virtualization.
Finally, my complements to CCICE for hosting the event, and to Isabelle Guis who acted as moderator. This was a great event, and a good opportunity for local entrepreneurs to get more involved with companies already engaged in building the new SDN ecosystem. And, one thing that became increasingly clear to me as I listened to the diversity of solutions offered by the panelists – the changes coming in SDN are just beginning. Further, the companies represented last night were offering very diverse solutions and in most cases were not really competitors (with the obvious exception of Cisco, who ultimately competes with them all!). There will clearly be many players here and much excitement. With the changes driven by virtualization and now SDN we are in for many exciting changes over the next several years, and I find this very exciting.
A very well balanced and well written blog. I found it informative and like the tone you’ve adopted which doesn’t contribute to the hype on this topic. You may want to adopt SEO and tagging mechanisms that will provide greater visibility to what you write, as your blog site and blogs do not come up easily in searches etc.