Sunday, June 3, 2012

SDN is Marketing to Wants, not Needs

There have been plenty of articles written about how Software Defined Networks (SDN) don't really solve a specific use-case that couldn't (actually or potentially) be solved with networking solutions that exist on the market today.  These are typically written by people that either have a vested interest in existing models of networking (eg. hardware vendors) or consultants that have existing projects and architectures in place (eg. complexity extends the contract). But what you don't often see is articles written by users (or "customers") that are clamoring for more of the same. When those groups show up at public forums or events, you almost always hear them talking about how they want their network to operate.
  • They want less complexity in deploying policies
  • They want less complexity in managing upgrades to 100s or 1000s of devices
  • They want new flexibility in detailing with complex situations like "BYOD" (or "CoIT", or whatever you can it when users want to bring in their iPad to work)
[NOTE: There are several other wants that are frequently mentioned; buy commodity hardware, invented new routing protocols, etc.. but they are typically the domain of the largest web operators, not on the immediate wish-list of Enterprise or Mid-Sized companies for the next 3-5 years]

Part of the reason you don't hear "the killer use-case" coming forward is because that isn't the overwhelming want. And as much as people can debate wants vs. needs, the reality is that with new things wants are often the driving force. When people make the choice to try something new, they typically don't do it so they can extend the past. 

So be careful when looking at SDN. The earliest adopters might not be looking to change because of an unmet need, but rather because of an unmet want

1 comment:

  1. So, I think you have to be a bit careful here. First, the line is a bit fuzzy--is simpler operations or better app/infrastructure interaction a want or a need. I think you can argue in both sides of that. Second, I think you have to be a bit more granular around the discussion--SDN is an ambiguous term--it means different things to different folks. I agree that in the enterprise the killer app still has not made an appearance, but in other segments like SP, there are some compelling use cases.

    But in general, i agree that folks should be wary of the bright-shiny-object aspect of SDN.