This week the SDN buzz was about the new executive director for the OpenDaylight Project (ODP). The Linux Foundation announced Tuesday that former VMware executive Nicolas "Neela" Jacques is now the first ODP Executive Director. Jacques hopes to serve as a facilitator of the project, and as a member of ODP, Plexxi thinks this will only benefit the project and accelerate the results. We look forward to seeing what Jacques contributes to this role. In other news, Plexxi has a new channel on SDNCentral, which you can follow for the latest news and videos from us. In Plexxi's own featured video this week, Dan Backman explains what kind of software is still resident on the physical switches and what capabilities are included in that software. Here is the video of the week and a few of my reads in the Plexxi Pulse – enjoy!
Michele Chubirka describes the return of the IT “generalist” for Network Computing this week. As an experienced IT professional, Michele reflects that specialization in IT has increased dramatically in the past decade. Michele suggests specialization hasn’t benefited the industry or IT work environments, but with information easily accessible today, it may not be necessary to focus so narrowly anymore. I think if you believe in a highly orchestrated future, the question is whether the people who build and manage that middleware are generalists because they touch everything, or if they are integration specialists because their specific skill is pulling things together. Whatever the answer, it's probably all nuance. The real implication of what Michele is discussing is that the changes vendors are pushing and customers are clamoring for extend well beyond technology. Those companies who recognize this as a real Change Initiative (capital 'C', capital 'I') are more likely to succeed on the first try. When you are dealing with peoples' lives, it goes well beyond technology. Maybe what we really need is a different kind of human specialist.
FierceEnterprise Communications’ Derek C. Slater explains that according to Nick Lippis, co-founder of the Open Networking User Group, SDN use cases won’t be available until 2015. People should be aware that deploying SDN is not like buying a new iPad. Companies who wait until 2015 to start making plans will be disappointed to find that they lack the onboard skills, organizational structure, appropriate metrics, planning processes, and even underlying architecture to support SDN. SDN is a transformational change that requires more than just turning some protocols on. There are in fact things that IT leaders can do today and considering the eventual end goal when making today's plans will make the transition significantly easier.
Sean Michael Kerner at Enterprise Networking Planet wrote about the new executive director for the OpenDaylight Project, Nicolas Jacques. According to Jacques the new position is more of a facilitator than an executive. I like the addition of someone who has vendor experience. The ultimate success of ODP will be its ability to get working code into production environments. In that regard, ODP is at once an open source project and a product. Understanding how to cultivate a developer community and deliver something that is commercial grade is really the overarching requirement. ODP won't be able to endure a slow burn, taking years to emerge. SDN needs to move faster than that, and if ODP isn't ready, other controllers will take that role.
Light Reading’s Carol Wilson says the networking market has been flooded with new vendors offering better solutions for data centers, which emphasizes the need for a standards-based open solution. The OpenDaylight Project is trying to provide just that. I think where ODP can succeed is in delivering a platform, because I believe ODP is more of a platform than a controller. If it really is extensible, then even if proprietary elements emerge, customers can still settle on a single point of control solution. Additionally, I think the reality in the short term is that people need to experiment to see what technologies will emerge. We should be pushing new code out, testing it, iterating, and failing or promoting quickly. Standards might be the outcome once we know more, but asking for consensus before you try something out seems a bit inhibiting.
David Weldon at FierceCIO discusses a new report from Dell'Oro Group that predicts the SDN market will increase by six-fold in the next five years, and continued adoption of cloud computing will be the primary driver of that growth. We should be careful treating SDN as a market unto itself. The reality is that SDN spend is just a subset of networking spend. It is unlikely to add net-new IT spend but rather substitute expenses that otherwise would have wound up in the general networking bucket. The impact to the larger networking market, however, ought to be significant. The number of RFPs today that require at least road-mapped SDN support is large and growing. As vendors rotate portfolios around SDN, SDN should continue to grow in terms of purchase impact.
A CloudTweaks staff writer explains why SDN and software-defined data centers (SDDC) were such a focal point at VMworld in both San Francisco and Barcelona. Because most enterprises have migrated to cloud and/or a virtualized infrastructure, the next logical step is deploying SDN. I agree with the premise of this article – that operational considerations will ultimately determine how and when SDN moves from labs into production environments, but I don’t agree with the comments about how people are not thinking through the operational considerations or putting plans in place to make applications better. VMWare's Martin Casado has posted openly about the need for troubleshooting and monitoring tools in a virtual environment. Cisco's announcements last week around the Application-Centric Infrastructure were certainly about orchestration and operationalization, and Plexxi is all about application visibility and performance. The fact that these topics have not hit the mainstream is concerning, but I don't think it's because they aren't being worked. It's more that the customer base needs to be first educated on bigger ideas. There will be a natural progression into operational considerations. Remember that networking sales cycles are long, and it takes people a while to get to the “nitty gritty.”