While Solid-State Disks (SSDs) are a well-known, effective and extremely expensive technique for speeding up storage performance, their applicability has been limited. Now a new generation of SAN-aware SSDs promises to bring their performance improvements to the SAN as a whole.
A solid-state disk is a large chunk of solid-state memory that is configured and treated as if it were a storage device like a disk. Usually SSDs are separate boxes with their own battery-backed power supply to give them the reliability of a conventional disk. The result is something that looks to the system like an extremely fast storage device. SSDs have been used for years to do jobs such as improving the performance of transactional databases. They haven't been common because they are so expensive. Traditionally a solid-state disk is attached to a single application, such as a database. In effect it becomes a very high speed cache for that application. But because SSDs are so expensive they are usually solutions of last resort and generally used only when a single application on the system needs an enormous performance boost.
The latest wrinkle in SSDs is to virtualize them as a SAN resource. This lets the
SAN-aware SSDs aren't any cheaper, but because they can be shared over the whole SAN, they can be much more cost effective. Instead of improving the performance of a single application, the SSD increases the performance of the whole SAN.
Imperial Technology has a discussion of the basics of using SSDs with SANs in a white paper titled "Tuning Storage Area Networks (SANs) with Solid State Disks" at its website. Another white paper on the same topic from Imperial and FalconStor Software titled "The High Performance Alliance: SAN, SSD and Virtualization" can be found at www.sanaccelerator.com/
Rick Cook has been writing about mass storage since the days when the term meant an 80K floppy disk. The computers he learned on used ferrite cores and magnetic drums. For the last twenty years he has been a freelance writer specializing in storage and other computer issues.
This was first published in August 2002