Are most problems hardware- or software- related?

Burmeister: Both. It can be multiple things going on, where more than one hard drive failed in a RAID system. That is something we see on a regular basis. It can be things like a RAID controller failing. The drives may be working, but there's no intelligence anymore that knows how that bank was put together. The third thing is what we refer to as human-error problems. When there are drive failures ... the process is to swap out the bad drive. People pull out the wrong drive by accident and get the drives in the wrong order, which results in a scrambling of data.

If my server hard disk crashes, wouldn't I be better off recovering from tape than paying a premium to recover data from the failed disks?

Burmeister: We've done a lot of work on server recoveries where we can rebuild a system faster than [IT] can restore a tape. A day of being behind with data [can be] worth millions of dollars.

What is it going to cost me to get that data back?