Unless you’ve been hiding in a cave for the past four years, you’ve probably noticed that virtualization is one of the hottest IT trends — for good reason. Virtualization is one of those rare technologies that almost doesn’t seem to require tradeoffs: It reduces costs, simplifies operations and improves performance — all at the same time.
Specifically, server virtualization reduces hardware costs by enabling consolidation, leading to fewer physical boxes. It reduces energy costs for the same reason (fewer servers in use means lower power consumption). It also enables users to standardize server provisioning and delivery, and reduce the time it takes to provision a server from weeks to hours.
Not surprisingly, server virtualization’s become almost ubiquitous in the organizations I work with — 93% say they’re currently deploying server virtualization, and 78% of them say they’re using server virtualization for customer-facing applications. Overall, 38% of workloads are virtualized.
But the one thing many IT professionals say they’re not using virtualization for is communications services — particularly unified communications. Why?
Certainly vendor balkiness at supporting virtual platforms has been part of it. Ensuring the stability of unified communications — with their integrated voice and data services — is challenging enough without introducing another layer of complexity and resource contention to the mix.
And, if vendors aren’t proactively enabling virtualized communications, IT folks are largely hesitant to get out in front of them and do it unsupported. One notable exception, though, was Exchange servers: A handful of folks who’d gotten fed up with the gymnastics of maintenance and with the slow pace at which Microsoft was moving to support virtual platforms (other than their own) took the bit in their teeth and moved Exchange onto VMware before it was an officially supported platform. The result? More stable Exchange servers, with all the benefits of virtualization—especially improved disaster recovery. (Redmond, are you listening? Faster off the mark next time!)
Another reason IT practitioners are slow to roll out virtualized UC servers is a concern with resource contention. After all, when you virtualize, you do it to consolidate workloads, right? And if your voice and e-mail servers are already chewing through the majority of available resources, why would you want to combine them with anything else?
Not so fast. One reason to virtualize is to consolidate, of course, and as noted, this can create some compelling ROI. But other reasons to virtualize include improving availability, disaster recovery and maintainability. And you can reap these benefits without necessarily using virtualization to consolidate.
Specifically, there’s a mode of virtualization – we call it one-to-one virtualization – that gets you a host of benefits without forcing your phone switch to fight for resources with some inventory management system or Web server. When the folks we worked with virtualized one UC system onto a host – no sharing – they got the benefit of hardware abstraction for maintainability and system availability. In other words, they could that instance move to a new server if the hardware started going bad. They also got improved disaster recovery (since storage replication handles availability).
UC vendors need to step up to the plate and start explicitly supporting virtual servers as platforms. But IT folks don’t need to wait until that happens to start reaping the rewards of virtualization. In particular, look into the possibility of using one-to-one virtualization under your voice and other UC servers to improve their manageability and maintainability.
