Solid State Drives – the promise of a free lunch?
I read an interesting article a few weeks ago on the SQL Server Magazine web-site where the issue of Solid State Drives (SSD) and their potential to impact the future need to tune databases was being discussed.
The article raised the question that as SSD becomes more mainstream, and its capacity increases significantly, then could it eventually eliminate the need for database designers/administrators to have to optimise table structures to deliver acceptable levels of performance?
The argument used was along the lines that with SSD there’s less traditional disk i/o going on (making reads a thousand times quicker than hard disks), so query performance levels may just be acceptable by virtue of the SSD memory delivering data quickly to the consumer process. This makes good sense, but also reminds me of previous technology advances in this area such as RAM disks and even paging files, which all promised such things but eventually needed cleverer system infrastructure around them to fulfil an overall business need.
That said, I have absolutely no doubt that SSD will make a significant impact on data storage access times (it has to). However, my guess is that it will just push the problem elsewhere. So, as much as we developers & technicians would like to think that it may deliver us a ‘free lunch’, I would suggest it’s more likely to be a ‘free starter’ and that (sadly) we will continue to have much more work to do to produce the main course and dessert that will keep our customers happy and coming back for more…