The big benefit of this is that the idea of if you do this then the vendors could be on a faster innovation curve. If you look at the compute world, there have been a lot of complaints about the fact that it has moved to a much more programmable environment in the last decade and networking hasn't kept up because it has been locked down in the firmware. As such, you don't have access.
There is value there because networks are different. There is a dynamic behaviour. But having said that, there are opportunities for networking to become much more flexible and that is the promise. Whether you are utilising that flexibility yourself or the vendor is utilising that flexibility, at the end of the day you get more capabilities, much more quickly.
It does seem that if OpenFlow takes off as a technology then the networking industry could experienced for massive disruption. Could a start-up come in and take the market away from you and Cisco?
Look at the wireless LAN market 10 years ago. We had what we called big fat access points. And then there was a big movement to make lighter access points and move all the intelligence to the controller.
It hasn't changed anything for the wireless industry, you are paying less for the access points and more for the controller, but the solution still costs the same. You have a lot more flexibility and capability, but you are not buying access points form one vendor and a controller from another. While in theory it is possible to do that, there are other factors when people buy a network. Do you really want to do all the work of integrating this equipment?
