According to performance management systems developer SevOne, nine out of ten enterprise executives say they are not confident in their ability to identify network issues before they impact end users. And nearly half say they encounter up to five issues per month, with a large portion of the remainder saying they simply don’t know how many problems they have to deal with on a monthly basis. On average, it takes about five hours to detect the problem, isolate the cause and effect a repair.
Clearly, then, visibility into disparate infrastructures is lacking. But the issue remains: how can you keep track of network goings on when you don’t actually own the network?
According to Riverbed, one of the keys is the ability to consolidate views across physical, virtual and cloud infrastructure into a single user interface. The company’s new Cascade 10.0 system works closely with VMware’s VXLAN SDN (software-defined network) architecture as well as the View VDI platform to track how applications are faring across multiple networks. The system brings VDI and CIFS monitoring and troubleshooting tools to the Steelhead and Granite optimization appliances, providing a unified management stack for virtual server, desktop and network environments. At the same time, the new Cascade Profiler Virtual Edition, Cascade Gateway Virtual Edition and the Cascade Shark Virtual Edition prove deep-dive visibility and performance management functions when launched within vSphere virtual machines.
In many ways, the fact that networks themselves are finding their way onto the virtual tier makes management and visibility a lot easier. Vyatta’s vPlane, for example, is a Layer 3 forwarding plane designed to foster software-defined virtual networks. As ENP’s Matt Sarrel explains, one of the key benefits is increased visibility because network architectures can be configured and reconfigured almost at will. This allows network administrators to not only see how the network is functioning on a granular level, but to quickly correct performance issues and streamline network pathways as data and application needs change.
But since many knowledge workers are already well-versed in wireless technologies, the boundaries between the two worlds are quickly falling apart in the enterprise. This has upped the urgency for a unified management stack, which Cisco says it has finally achieved with the new Prime platform. The system provides a single user interface for wired and wireless networks, and concentrates mostly on maintaining application performance rather than network resources themselves. A key function is sub second stateful failover in the company’s wireless controllers, so if a unit does go bad the system reroutes traffic to a new controller almost instantaneously.
The very idea that we can even be talking about a unified network architecture that covers everything from old-style routers and switches to the latest virtual and mobile systems is a technological marvel.
In the near future, hardware will still fail, but pathways will shift on the fly, and technicians will be able to swap out the bad component at their leisure, rather than in a frantic rush because the entire data universe has come crashing down.