As expected, the two SSDs in RAID 0 had an easy time outclassing an array of up to eight 3.5” SAS hard drives that spin at 15 000 RPM. The SSDs are better in almost every benchmark with the exception of throughput, which typically scales nicely with the number of drives used. This is where the hard drives won through sheer force of numbers.
It is important to note that there are faster and somewhat more efficient hard drives available, which would have shifted the results a bit in favor of the hard drive arrays. Yet, the differences we see are significant enough to be sure that whatever you tweak on the hard drive side still cannot beat a fast and efficient SSD array. Also, we had to use the 15K RPM Fujitsu drives because no hard drive maker wanted to provide drives once we told them what we were planning. This alone should speak volumes.
Although we used a set of SSDs based on the latest technology, which introduces DDR technology into the SSD world, the results can be considered representative for modern SSDs in general. We saw up to 12x faster performance in I/O operations, and power efficiency can increase many times if you look at it on a system level—even more if you simply compare data on a drive level.
In the end, you should decide whether it’s time to adopt SSDs over HDDs by examining the scenarios in which hard drives still make sense. Mission critical systems may have strict validation requirements. There could be cost or capacity conditions placed on a configuration that take precedence over I/O performance.
From a cost point of view, $699 for a high-end 256 GB SSD is still a lot compared to less than $300 for a very fast 15 000 RPM, 300 GB SAS hard drive. But $399 is reasonable for such a fast 128 GB SSD. However, once you start consolidating storage systems, recognize the long-term power savings, and are able to get potentially more work done in far less time, owning a smaller number of SSD-centric machines start to get seriously attractive.