Posted by Mae Kowalke on Tuesday, May 26, 2015 with No comments
Memorial Day is over, and it's back to business as usual--which includes pondering a number of factors related to deploying Software-Defined Networking (SDN).
For one thing, there's a lot of buzz about not only SDN but "software-defined everything" or SDx. SDxCentral pointed out that for most of networking's history, changes came mostly through hardware-based innovations. Now, software is driving adaption, forces dramatic changes.
This trend toward 'digital' is driven by the expectations of consumer ("I want what I want, and I want it now!"), globalization and competition, and the evolution of the cloud, among other influences.
But just how limitless is the SDN concept? With some additional evolution, SDxCentral predicts SDN is just the beginning and SDx is limited only by our imagination. Sure, compute, storage, and network are IT hardware pillars being changed by SDN. But what else in the IT world could become software-defined?
SDxCentral thinks some of the possibilities include security (achieved through deep vendor integration via APIs), operations (achieved through a devotion to automation), and management (the piece that brings everything together in a 'single pane of glass.')
Turning these concepts into reality requires next-generation IT infrastructure (aka SDxI) designed for an SDx world, notes SDxCentral. SDx puts a heavy tech burden on networking especially, requiring flexibility, personalization, security, efficiency to scale.
The good news, SDxCentral said, is that just about every part of networking infrastructure now has a software based component, including Network Functions Virtualization (NFV), security, storage, data center, and WAN.
Getting traction for all these software-defined things, though, requires more development of use case definitions and product architectures, so that software and hardware can be appropriately integrated and orchestrated.
It's now clear, SDxCentral concluded, that typical SDN deployments are evolutionary rather than revolutionary: traditionally hardware-based network infrastructure will be gradually replaced with software-defined equivalents. Assuming interoperability and deployability are accounted for, budgets and refresh cycles are the focus.
What all this means is that many SDN deployments are likely in the next 3-4 years, as shorter refresh cycles in the real world drive companies to seek new ways of reducing total cost of ownership (TCO).
What other SDN deployment considerations are, or should be, top of mind?