Technological advancement is a journey, not a destination. For all the talk about building a software defined network, we should remember that SDN is likely to be a work in progress for many years to come, and in the end the typical enterprise is probably not going to wind up with what was first envisioned.
This is not a bad thing, however, because if SDN does anything for data infrastructure, it introduces a level of flexibility and architectural dynamism that will enhance the ability to adapt to changing data conditions.
But how do we get there from here? And more importantly, if we don’t know what total success in SDN looks like yet, how are we to avoid abject failure?
According to IDC, the next two years will be crucial in the rollout of SDN. The company is expecting the market to grow from $960 million today to more than $8 billion by 2018, with much of the growth fueled by the traditional enterprise community, as opposed to the nascent cloud industry. Much of this activity will likely center around greenfield deployments, perhaps of the modular variety, but there will also be a fair amount of legacy infrastructure conversion and all the inherent complexity and conflicting decision points that go with it.
The thing to keep in mind about SDN, says Big Technology UK’s Jason Dance, is that it is not simply a matter of replacing physical assets with virtual ones as we saw in the server farm. Rather, this is a complete reworking of the network architecture from the north-south orientation of the edge-core configuration to the east-west style of clustered fabric topologies. Not only does this rewrite the economics of networks and network infrastructure, it dramatically increases the enterprise's ability to adjust its digital footing to respond to and capitalize on changing market conditions. The danger lies in implementing this transition in such a way that it jeopardizes existing application environments.
This is why an incremental approach to SDN is warranted in most cases, says Gartner’s Andrew Lerner. Some may criticize the use of overlays as “SDN lite,” but the fact is that legacy infrastructure still handles the vast majority of data traffic, and real-world experience with SDN is hard to come by. Again, though, a word of caution: improper implementation that does not take legacy design elements into consideration can degrade performance and ultimately hamper the change-over to full SDN. And be sure your plans for SDN encompass more than just network virtualization to include things like staffing, management and visibility.
A good way to go about this is to employ Abraham Maslow’s Hierarchy of Needs, says ITC’s Mike Fratto. Devised as a method to parse out ways to motivate people to achieve their goals, it can provide valuable insight into the motivations behind SDN deployment and other technology initiatives and whether the solutions being offered are really in line with organizational needs. Before you even contact a vendor, then, it helps to conduct a clear-eyed assessment of what SDN does and does not do, and then make sure all stakeholders are behind the dramatic infrastructural and architectural changes that are about to take place.
Clearly, the stakes are high with SDN. Networking is such an integral part of the data experiences that problems associated with major systems upgrades are not easily ignored.
Nevertheless, change is coming. The only thing IT executives can do is plan for and try to manage it to the enterprise’s best advantage. And by all means, don’t think you can simply ignore it.
Photo courtesy of Shutterstock.