ONS 2013 Thoughts…
I had every intention of producing several long posts about ONS 2013, but events in my hometown coupled with a busy meeting schedule at ONS resulted in not finding a lot of time to focus on writing. I think my colleague Mike Bushong summed up much of my thoughts here and I would like to add a few other thoughts I have about SDN after ONS 2013.
1. It seems ONS 2013 had a common theme around the need to for many people to articulate their definition of SDN. There were many speakers quite willing tell us what SDN is, where it came from, what it is not, what it will or will not do and so forth. Personally, I really do not care if we agree or do not agree on the definition of SDN. I think SDN is about the implementation of a controller architecture. Maybe that is an open source controller, maybe that is nested controllers, maybe that is private controllers, overly controllers or hybrid controllers and maybe they speak OpenFlow or not. My points are: (i) can your SDN system construct a network using a controller architecture, (ii) what can this network do and (iii) does this controller based network deliver better value than traditional switched networks based on distributed protocols? Those three points are the only points that matter in the business of SDN. That leads me to my second thought post ONS.
2. It is apparent to me as a participate in the networking industry that 2013 is moving day, to use a golfing analogy. There has been a big venture capital exit ($1.2B) by the Nicira team in 2012. There are sizable venture funded startups: Plexxi ($48.5M), Big Switch ($45.3M), Pluribus ($61.5M), Pica8 ($6.6M), Cumulus Networks ($?); a number of spin-in type companies Nuage, Insieme ($100M), Contrail ($176M acquisition by Juniper) as well as a whole host of public company initiatives (CSCO, INTC, HPQ, DELL, NEC) as well as various consortiums of Open Networking Foundation (ONF), OpenDaylight Project (ODP), Object Management Group (OMG) and not to mention the efforts of Google and about 20 other companies at ONS 2013 demonstrating something SDN related.
Ultimately, I am not an engineering person. I collaborate with vastly smarter people who work on the technology details of Plexxi. I do consider myself to be knowledgeable about building companies, specifically technology startups. Everything around me in the networking or SDN space tells me it is moving day; the year to be be winning networks. The winners in the SDN space are going to be selected by the end-users. That is the crux of the Cisco vs VMware debate at ONS 2013. It is about winning market share and market share is decided by the customer. Influencing and winning customer mindshare is a step to controlling marketshare. Everything else is meaningless.