Virtualization: The future is in the fabric

Storage Networking World - Virtualization - it's just a word. But in storage circles, this word can evoke a wide range of emotions, from hyper-excitement over the endless possibilities, to downright annoyance over its true meaning.
One way or the other, storage virtualization is inevitable. In fact, the fabric-based incarnation of storage virtualization that we are now witnessing is just another inflection point on a continuum that had its beginnings when IBM transitioned mainframes from System/360 to System/370 back in the late '60s.

It's an engine
Storage virtualization engines - hardware/software processes that present logical representations of physical storage - have lived happily, and anonymously, under the mainframe covers for many years. Few if any application users have even been aware of their existence.
What is now happening is that these virtualization workhorses are following a course that mimics the progress of computing itself: from centralized to distributed, from proprietary to open, and from high-priced to commodity.
They've moved from IBM's Dynamic Address Translation (DAT) boxes of 30 years ago to distributed, open systems host environments (Veritas' Volume Manager for Unix, Sistina's VM for Linux). From there, it was on to storage arrays (STK's SVA, XIOtech's Magnitude and many others), and now to appliance and switching platforms embedded within the storage fabric.

The future is in the fabric
So what's the big deal? Why are vendors cranking up the hype generators once again? Because a virtualization engine spinning away within an intelligent switch has the very significant potential to move existing storage and data management applications to the nexus of the storage fabric, and to create new apps as well. Here's a sampling of management applications now destined to find embedded homes within the storage fabric:

• Replication and migration processes that copy and move data around the storage fabric without regard to proprietary storage array dependencies.
  
• Backup and recovery applications that operate independently of LANs and application servers.
  
• Data lifecycle management applications that look at data from the perspective of usage patterns and value to the enterprise.
  
• Automated, nondisruptive storage provisioning tools that provide additional storage capacity to user applications as needed and take capacity away from those with too much.

Reprinted with permission from

This article is reprinted by permission from SNW Online.
Story copyright 2006 SNW Online, all rights reserved.