Ronald Hudson - Fotolia
The price of an array isn't the only factor organizations looking at adding flash storage should consider.
The capital expense of purchasing solid-state drives is an important part of determining the true cost of flash, but it doesn't paint the whole picture. Other aspects -- such as the cost of IOPS -- can help determine whether flash is a good investment. For example, it's possible to achieve high numbers of IOPS by striping together hard disk drives, but it could take an excessive number of HDDs to reach the desired level of performance. Flash drives deliver a much higher number of IOPS than HDDs, so it is possible that using flash storage may reduce overall costs, even if the cost of flash drives is higher than HDDs.
One reason flash storage has a reputation for being expensive is that flash drives have relatively low capacity, so the cost per gigabyte is high. But using HDDs for high-performance workloads can actually cost more because you need more of them and could end up with wasted capacity.
HDDs tend to have a relatively high capacity, and if you stripe many together to meet a workload's performance requirements, the resulting volume will have a lot of capacity. That's fine if you actually need all of that space, but think about the cost implications if your workload only requires a small fraction of the volume's storage. If you only look at the cost per raw gigabyte of storage, then high-performance, high-capacity HDD volumes have a low cost per gigabyte. But if you look at the cost per gigabyte of the storage that your workload will actually use, the cost per gigabyte is much higher. The latter comparison makes HDD and flash storage analogous, especially for high-performance workloads.
Consider refresh cycles
Another concern when determining the cost of flash storage is the hardware refresh cycle. Most organizations replace their storage arrays every five to seven years. But the drives within those arrays may need to be replaced more frequently depending upon how often users access the data stored on them. Early on, flash storage had a reputation for having a short lifespan because the NAND chips wore out quickly. Over time, however, manufacturers have significantly improved flash storage durability. Not only have the NAND chips themselves improved, but technologies such as wear leveling help to extend the life of the drives by distributing write operations evenly across the available cells. You can sometimes reduce costs further -- and improve disk longevity -- by using data reduction techniques to reduce the number of write operations.
Consumer vs. enterprise flash
Storage vendors commonly offer both consumer-grade flash storage and enterprise-grade flash storage. There is sometimes a subtle difference in performance, but the most significant differences between the two are cost and longevity. The cost of flash in enterprise-grade drives is higher than in consumer storage, but also tends to have a much longer lifespan.
When evaluating the ROI and cost of flash storage, it is important to consider which type of flash will be most cost effective for the workloads you're running. Enterprise-class flash costs more, but it might be the less expensive option in the long run if it means you can replace disks less frequently. This is especially true when you take the labor costs associated with storage maintenance into account. Enterprise-class flash sometimes includes lifespan monitoring tools, whereas consumer flash will leave you guessing. With consumer flash, you could incur additional costs as a result of replacing disks more -- or less -- frequently than necessary.
The products AFA vendors offer
What sets AFA vendors apart?
For total cost of ownership, the advantage goes to all-flash arrays over HDDs