Wednesday, 4 August 2010

Optimizing the Virtual Cloud

BURBANK, CA–(Marketwire – July 29, 2010) – Cloud computing is a revolution for corporate data systems. Instead of having to install and maintain costly server hardware on-site, enterprises can now subscribe to a cloud service and literally use computing resources as they are needed.

When they are no longer required, those resources are used elsewhere. The idea has firmly taken hold; a prominent motion-picture company recently turned to a cloud for their highly compute-intensive animation needs, and certainly saved major costs by not having to utilize in-house resources.
One element that makes cloud computing so attractive is its use of virtual servers. It boggles the mind to think that a user at a corporation can access the cloud and actually launch their very own complete server for needed tasks. When that task is done, that server is no longer needed and it simply ceases to exist. The incredible gains that virtualization has brought in hardware and space economics — to cloud computing and everywhere else — are still being realized.

But as we all know, while we all have our heads in the clouds it is important to keep our eyes on the ground as well. In the case of cloud computing, that "ground" is the hardware hosting virtual systems — hardware that is still subject to drastic system slows courtesy of file fragmentation.

All hard drives suffer from file fragmentation — and hard drives are the place that data for virtual systems, and hence a computing cloud, is stored. In comparison to a traditional server, a virtual environment has a few added steps to data storage and retrieval, however, and fragmentation can have even more of an impact. When a file request occurs on a virtual server, the I/O request is relayed, at the least, from the guest system to the host system — which means multiple requests are occurring for each file request. When a file is fragmented into hundreds or thousands of fragments, there are multiple I/O requests for each fragment. This operation creates an enormous amount of unnecessary overhead on disk subsystems.

Virtual disks also suffer from "bloat" — their sizes are dynamically set to grow, but they don't shrink when users or applications remove data. This wastes the space that could be allocated to other virtual systems.
The latest in technology automatically and invisibly prevents a majority of fragmentation before it occurs, totally negating the effects of fragmentation in virtual environments. Because free space is also consolidated as part of the process, virtual disk "bloat" is eliminated. Enterprises can now take full advantage of cloud and virtual computing without ever having to worry about the performance drain from fragmentation again.

No comments: