Software-Defined Networking (SDN) seems like a dream come true for the folks who run and manage networks: LANs, WANs, and the like. Makes you wonder why this movement has been so long in the making.
So far most of the SDN action appears to have been taking place on the service provider side. Adoption in the enterprise is still understandably slow. Clearly it is not easy to rip out /make wholesale changes to current data center environments.
But the move to SDN is expected to accelerate; there is so much to gain. End of the day we should see a software centric network world that is (a) based on commodity hardware and (b) controlled entirely through applications. This should drive far reaching changes in terms of retooling and reskilling in the marketplace. Unfortunately many of the current networking skills seem likely to go out of fashion.
The SDN controller will emerge at the core of the network. Basically this application (yes it is software) provides a southbound interface to control traffic across devices and a northbound interface for higher level apps to deliver specific, useful functionality for the network.
Clearly the northbound interface is a game changer with the APIs to a wide variety of SDN controllers that abstract network functionality, presenting a centralized view to a higher applications layer (whether the seven layer OSI model will continue to be applicable in its current form is perhaps a question to be pondered).
HP has announced an app store for SDN apps albeit one that seems to work only with their SDN controller. Yet they have taken the lead here in what looks like a promising step that could hold the key to the future.
Today there are several SDN controllers available – commercial and open source ones – that don’t really talk to each other much.
As in other scenarios, a shakeout therefore appears to be inevitable. With a set of providers left standing, each with their own SDN app marketplaces that support their respective SDN controller products. That in turn should fuel a robust SDN apps ecosystem that presents an interesting parallel with the current mobile apps market.
It will be fascinating to see how Cisco manages this transition.
Will they continue to be the undisputed leader? Looks unlikely, especially in the long run. Will they continue to be a leading player, one among several? That does seem plausible.
One thing is for sure. Exciting times lie ahead for the world of packets and frames. Where they go will no longer be the decision of the solitary switch. Instead it will be an array of apps pulling the strings from afar.