I wrote a commentary for the SQL Server Magazine e-newsletter last week, and was expecting a to get a bit more feedback than I did.So I'm hoping my blog readers can help out. I was commenting on a message that one of my business partners had sent out wondering about the changes that new higher speed storage technology might bring.
Please read the whole commentary here (It's not that long.)
Basically, the comment is that new storage technology is appearing that could have I/O's per second (IOPS) rates thousand of times faster than they are now.
Most of the reasons for building good indexes and tuning our queries is to reduce the number of reads that SQL Server would have to perform. But if the reads are so fast that even with tables scans our queries are always 'fast enough', would you spend time on query and index tuning?
At this point, my question is NOT whether you think there will be available and affordable storage systems that can perform as suggested. My question is "IF there were such systems, would you still spend time tuning your queries and evaluating the best indexes?" Also, if query tuning does become unnecessary, what would you do instead to maintain your database system's performance? Or would you be out of a job?
I'd really like to hear what you think!