With the advent of low-cost Serial ATA (SATA) disk drives, many organizations have rushed to implement this technology for all but the most demanding storage requirements.
Richard Kruszewski, network administrator at TBWA/Chiat/Day, Los Angeles, says his organization was able to switch most of its storage to SATA, saving money and achieving acceptable reliability in the bargain. Kruszewski had started out looking for a primary storage solution to serve over 500 users in his company's Los Angeles office. These users needed regular, fast access on a daily basis to a variety of data types ranging from simple Microsoft Word and PowerPoint files, to high-resolution images and video. After considering Fibre Channel disks, Kruszewski began to evaluate SATA and was impressed by the cost savings. In fact, SATA was such a success it was also adopted in the company's New York City offices.
The price point for SATA disks is tempting, but companies that are thinking of implementing SATA should develop a management plan beforehand, says senior analyst Brad O'Neil, Taneja Group, Hopkinton, Mass. The best and most cost-effective storage strategy, he explains, will appropriately match the demands for access and archiving with the capabilities of specific technologies.
In a recent Storage Networking World presentation he offered advice on the best way to incorporate low-cost disk into a tiered storage plan.
"One of the most obvious ways to drive up ROI is to do what I call platform-appropriate storage," he says. To accomplish this, O'Neil recommends a detailed audit of applications and their associated service level agreements, and then line up each application up with the appropriate storage technology.
He points out that some databases don't really need mission critical storage support. "When you have the opportunity to cut storage costs in half, you see a lot of companies that choose differently."
In the end, SATA isn't only about cutting costs: it's also about enhancing capabilities. O'Neil notes that many IT managers now want to keep as much information as possible online and available 24/7, and that means using disk.
For Kruszewski, after "losing" on SATA disk, he says he has made a point of implementing RAID-5 and mirroring. "That way if I lose a disk the users won't notice anything," he adds. "We wanted this to be highly available and scalable -- all the usual buzzwords -- and it is," he adds.
For more information:
About the author: Alan Earls is a freelance writer in Franklin, Mass.