InfoStor reports, in a recent article, that virtualization is being utilized by many top enterprises to maximize system resource ROI . It’s certainly no surprise; CIOs are always looking to strengthen that bottom line, and when multiple “computers” can be run from a single hardware platform, that bottom line can certainly be strengthened.
The article also takes up the major challenges that virtualization brings. For example, capacity planning, when dealing with complexity in mixed environments, becomes increasingly difficult. Resource tracking and utilization becomes a multi-headed hydra. Historical analysis becomes almost moot when prime resources are constantly shifting and being reallocated. Device failures, since they are utilized by multiple virtual machines, trigger multiple alerts and tracking down the actual failure becomes an interesting exercise.
While meeting these challenges, the last thing CIOs and IT staffs needs is a hidden, crippling disease affecting every single hard drive in a system. And yet, this is the case with file fragmentation–the saving of files in multiple pieces (fragments) in order to fully utilize disk space, a function of all operating systems. While a major performance problem on a single machine, fragmentation seriously compounds performance issues on multiple virtual machines all accessing the same drive or set of drives.
When a file is accessed on virtual machines, that request is, of course, passed along to the host system. Hence, multiple requests are occurring for each file request–minimally, one request for the guest system, then another for the host system. But in a common fragmentation scenario, especially with virtual servers creating high amounts of disk activity, files will be fragmented in to tens, hundreds or even thousands of fragments. Imagine the frantic activity with multiple I/Os for each fragment of each and every file requested. The impact on performance is horrendous, affecting IT service and employee production across the boards.
Most sites have implemented some sort of defragmentation solution. Many utilize scheduled defragmentation–a solution which can be schedule to run when server traffic is at a minimum and production won’t be affected. But there are hidden costs associated with scheduled defragmentation that, especially in today’s frantic computer environments, can reduce or even eliminate the ROI that defragmentation should provide. IT staff are having to take the time to analyze fragmentation–another severe challenge in a virtual environment–and worst of all scheduled defragmentation, since it only runs periodically, is actually leaving performance-crippling fragmentation behind.
The best possible solution, one which is only now becoming available, is a transparent background defragmentation solution which requires no scheduling and which has no impact on system resources. IT no longer need to spend time with it, and can continue to address problems and employ solutions to better service company personnel and customers. And best of all, such a solution defragments “on the fly” or as it occurs using only idle system resources, which means no more file fragmentation and an end to this hidden crippling disease.