These days, software defined networking is the talk of the town. As the final piece of the virtual data center puzzle, SDN stands as a defining moment in the transition from bricks-and-mortar enterprise data environments to fully interoperable, distributed cloud computing.
The changes that SDN will bring reside mostly on the architectural level. Lower on the network stack, however, some significant developments reflect not just the advent of new virtual computing environments, but mobile computing, Big Data analysis, and the fundamental shift in the way we humans interact with the digital universe.
Possibly most significantly, users continue to gain unprecedented control over data infrastructure, once exclusively IT's domain. Users may soon be able to scale resources at will, leaving the enterprise to either place artificial restrictions on utilization or establish an infinitely scalable sandbox – at home, in the cloud, or both – for employees, partners and even customers to play in. As cloud consultant Thoran Rodrigues noted recently, all cloud activity resides in a data center somewhere, so if you want the former to function properly, you pay more attention to the latter.
Indeed, the very notion of what is and is not the data center now looks a little fuzzy. This has a lot to do with how we view the network. In the old days, the enterprise network meant the LAN. Anything beyond the “edge” was the WAN. Nowadays, thanks to the increasing interoperability of colocation and cloud services with local data environments, the “data center” can easily reach across campuses, regions, or even nations. The rising demand for Application Delivery Controllers, which allow application functionality across disparate network environments, demonstrates this on a micro level. Traditional WAN systems, meanwhile, are already on the wane according to Infonetics, most likely due to the fact that cloud interoperability depends more on management and automation than on bandwidth, at least for the moment.
The rise of service-based provisioning will bring about a need for greater network intelligence as well, according to CommScope’s Charles Le Mahieu. Automated Infrastructure Management will play an increasingly vital role in the health and productivity of the data environment by feeding real-time connectivity information to both data and facilities management stacks. This will deliver not just lower costs and improved utilization, but better reliability and availability, not to mention an effective means to analyze network strengths and weaknesses when the front office is setting next year’s budget priorities.
You can also expect the network edge to become more sophisticated as the decade unfolds. With the individual data center acting as but one node in a distributed architecture, pushing data and services to the edge looks like the best way to prevent internal network infrastructure from succumbing to a barrage of requests. As Intel’s Diane Bryant points out, wireless base stations are already being upgraded with intelligence and computing power to reduce latency in service delivery.
So while it may seem that all the action these days is on the virtual layer, infrastructure feels the pressure to evolve as well. With the data center’s place in the digital universe less stable than before, the network needs a higher degree of flexibility to ensure it can connect stored data with processing resources and various client devices that are constantly on the move.