The advantages of solid state drives (SSDs) now seem to make them winners over hard disk drives (HDDs) in many computing applications. SSD vendors point to the ease of handling intensive storage input and output, the rapidity of making system snapshots and the resulting improvements available in disaster recovery objectives. Get RAID 1 performance, RAID 5 capacity and RAID 6 protection all-in-one, they claim. For IT managers who have been balancing these different goals together with the prospect of decreasing error protection with increasing RAID capacity, the flash drive solution may sound – almost – too good to be true.
It’s true that some SSD manufacturers are building in some very advanced capabilities into their products. Not only do their flash drives clone system memories in milliseconds, but they also use data block deduplication inline and as part of the process to store data even more efficiently. Data blocks that are repeated are identified and stored only once, with metadata pointers to reference different instances. Encryption at rest is also on offer, while still reducing bottlenecks that affect spinning storage systems trying to handle many parallel read and write requests.
So is it out with the old HDDs and in with the new SDDs? Not so fast is the answer. The SDD advantages may still cost more to achieve. While unit costs of data storage continue overall to fall, solid state drives are not following the trend in a consistent way. Historical pricing data shows that they were more expensive in 2014 than they were in 2012. Projections of current pricing trends suggest that even by 2020, SSDs for large-scale IT operations may still be at around $0.15 per gigabyte compared to an estimated $0.03 per gigabyte for HDDs. It looks like IT departments that want the superior performance will have to dig a little deeper into their pockets – but isn’t that always the case?