Introduction

Lately, the trend to double components to double performance seems to be affecting all computer components across the board from dual channel memory to dual GPUs with NVIDIA's SLI to dual core processors of tomorrow. One of the early segments of the computer world to take this approach is the storage area with RAID. RAID has been fairly ubiquitous in most higher end motherboards in the last few years but adoption on the desktop side, to me seems kind of low, despite the fact that there are obviously power hungry people out there who will not bat an eye to pick up a dual channel memory kit or a second board for SLI. But where does the aversion to RAID come from?

Cost -

Most boards on the consumer level support RAID0, RAID1 and RAID0+1. RAID implies a minimum of two hard drives (or 4 in the case of 0+1) which is a doubling or a quadrupling of the cost of storage.

Upgrading -

When do people upgrade hard drives? Usually when they break or when they run out of space. This will put RAID1 out of the equation since it will net zero gain in available disk space.

Complexity -

Computers are complex enough that reinstalling Windows is a major chore for many people. Setting up RAID is another step in an already complicated process.

Trade-offs

RAID0 increases the speed at the cost of data integrity. RAID1 provides redundancy over a single hard drive and RAID0. There are instances where RAID0 would be useful and instances where RAID1 would be useful but to apply RAID0 or RAID1 over an entire hard drive is not desirable in many cases. Personally this has been the biggest obstacle in RAID adoption - I value my data too much to run RAID0 but not quite enough to half my disk space.