When considering the potential benefits that typically drive server and desktop virtualization projects, questions about the underlying storage may never come up. They should. If we look at these benefits; maximizing hardware utilization, insulating the software environment from hardware configuration variables and ensuring that applications keep running, despite system outages, the state of the storage system is, indeed, of primary concern. It’s a fact that virtual infrastructures are completely dependent on the behavior of their underlying shared disks.
It’s well known how easily hypervisors migrate workloads from one physical server to another when one goes down. But what’s not as obvious is what happens when the shared storage device that both of those servers rely on is taken out of service. Taking storage systems down for a routine firmware upgrade, a water leak in the computer room, or just capacity expansion can cause major data outages for numerous virtual machines and the hundreds of desktops that they support. This is true regardless of how reliable a particular disk subsystem has proven to be. Fortunately, there are reasonable solutions available to eliminate this risk.
Companies such as DataCore offer the necessary storage virtualization software to:
a) maximize the utilization of existing storage devices,
b) insulate the software environment from hardware differences, and
c) ensure that applications keep running while storage devices are taken out of service.
Similar to server and desktop virtualization products, SANSymphony storage virtualization software abstracts most commonly used storage hardware devices into idealized, virtual disks. In the process, it automatically mirrors the information between physically segregated devices so that complete disk subsystems can be taken out of service without impacting applications.
Basically, this device-independent, virtualization technology allows normal changes that regularly occur in the storage infrastructure to go on without impacting the servers and desktops that depend on it. And storage virtualization software can repurpose existing disk systems attached to servers for use as part of a broader tiered storage pool. This enables the ‘mixing and matching’ of different models and different vendors, swapping these systems out or growing them in the background without affecting business processing.
The same software can also stretch the storage pool across a campus or a metropolitan area to provide additional safeguards against site problems, allowing it to be managed as one centralized resource. Look for built-in capabilities like cross-device snapshots, remote replication and thin provisioning to get the fullest use from all existing disk assets, despite their mutual incompatibilities.
Server and desktop virtualization initiatives promise significant benefits, like asset optimization, simplified operations and improved application uptime. However, the storage infrastructure that supports this compute environment must be up to the task. Layering in storage virtualization software at the outset as a crucial part of the overall virtualization strategy will ensure that you realize the benefits that first justified the move to a virtual infrastructure.