Database management systems (DBMS) put heavy demands on storage. If you have the opportunity to select disks specifically for a DBMS system, you can maximize your performance by making the right choices.
The faster the spindle speeds, the lower the disk latency. Go for the fastest RPM disks you can fit into your budget. Disk cache size is somewhat more controversial, because cached disks get their performance boost from the assumption of non-random data access. In a transaction processing system, for example, the conventional assumption is that the database is randomly accessed. In practice that might not be true, and your DBMS performance might benefit from the larger disk caches. Generally, if you have to choose between a faster disk and one with a bigger cache, go for the faster disk.
Because controllers can read and write to several disks at once, more disks mean higher performance. Fewer, larger disks may be cheaper, but more, smaller disks will give you better performance.
Several small arrays give better performance than one large array. Not only can you balance the load better, but multiple arrays allow you to use different storage strategies for various parts of the database. For example, your log files and database files can go on separate arrays to minimize read and write conflicts. RAID-1 can be used for your log files while the database files can be put on a RAID-10 (mirroring and striping) array.
Of course, this performance comes at the expense of complexity. Having multiple RAID levels for the parts of your DBMS will complicate maintenance and restoration. Try to strike a balance.
For more information:
About the author: Rick Cook has been writing about mass storage since the days when the term meant an 80 K 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.