As far as enterprise network architectures go, flat is the new black. And while the benefits of doing away with today's multi-layered infrastructure are clear both from deployment and management perspectives, it's also obvious that the journey will be neither simple nor cheap.
Nevertheless, the transition is growing more imminent by the day as the top networking vendors put the finishing touches on their converged network architectures. The latest, of course, is Cisco, which recently announced the latest upgrades to its Data Center Business Advantage portfolio that entail enhancements across the company's major switch and server lines as well as management and operating systems. Highlights include new Nexus 5000 devices featuring a unified port capability that allows for dynamic switching between 1 GbE, 10 GbE, FC and FCoE configuration, and a new Fabric Extender Technology (FEX) designed to integrate switches, servers, adapters and even VMs into a cohesive data environment.
One of the key benefits to this kind of integrated approach will be the reduction in actual networking hardware and infrastructure needed to support expanding virtual and cloud environments, according to InformationWeek's Charles Babcock. The unified port technology on the Nexus 5598, for instance, means you can alternate them between LAN and SAN operation without having to change physical cables. So as data loads dynamically shift from storage to communications, as they are doing with increasing regularity, network resources can accommodate them without having to build out maximum capacity across multiple layers.
According to Cisco's Canada rep, Jason Reil, the underlying concept here is shifting from a server-centric view of the data center to a hypervisor one. As he explained it to eChannelLine, the more easily a VM can interact with the physical network, the more mobile and scalable the entire data environment will be. And part of that is getting rid of patch panels and other layering obstacles that not only hinder flexibility but add service degradation and latency.
That's why the management aspect of the portfolio is so important, according to Paul D'Druz, Cisco's head of data center technical strategy in the UK. Enhancements like FCoE awareness at the switch level allows individual switches to alert its peers, as well as central management, when it is handling FCoE data. Not only does this offer SAN admins greater insight into FCoE behavior on the Ethernet fabric, but it affords greater control and functionality as data winds its way to and from the end points.
As I mentioned, this is ambitious stuff, and it doesn't exactly lend itself to plug-and-play deployment. But now that the pieces are in place, all that's left is the execution.