Skip to content

Is RAID a stopgap?

An article on setting up a software RAID on Linux got me thinking about nonvolatile mass storage technology advances. For the last 25 years or so, the rotating magnetic drive has been cheap enough for even personal computers as well as reasonably fast. Capacity increases have greatly increased their cost effectiveness and speed increases have contributed to the computing speed equation as well.

RAID was developed to provide reliability through redundancy for disk drives. The techniques also allowed some paralleling of data movement which could speed things up. Technology seems to be overtaking these goals in other ways which makes RAID seem like one of those technologies that may be on its way out to the esoterica farm.

Twenty years ago, a megabyte of volatile memory and a gigabyte of hard drive were cutting edge. Today, it is feasible to put 64 gigabytes of volatile memory and a few terabytes of hard drive in a PC. In addition to that, nonvolatile memory is showing up in solid state disks (SSD’s) that push the rotating magnetic media back on the stack towards library and archive use rather than towards the operating system front line. SSD’s with the newer SATA or even PCIe interfaces provide access speeds that exceed RAID capabilities and are getting to where they can fill the bus and overwhelm the CPU.

While speed is becoming a lost issue for RAID, redundancy is still there. It seems, though, that this can be handled in a variety of ways, often simpler ways than RAID.

Something to think about: where will RAID be in ten years?

Post a Comment

You must be logged in to post a comment.