Five Tips for Finding the Right SSD

Esther Spanjer is director, marketing management, SanDisk Corporation.

Esther_Spanjer_tn

Enterprise organizations face an interesting dilemma when it comes to meeting their data storage and application needs. In many cases, they are forced to balance the need for fast, reliable, consistent storage, with having access to limited resources. Often times this leads to IT managers selecting the lowest cost option, knowing that it fits within the budget and assuming that advertised speeds are representative of how the drive will perform in their environment – a win-win, right? Not so fast!

When selecting an SSD for your environment, there are a variety of factors that you need to take into account if you are going to eventually choose a solution that can properly meet your needs. Below are five tips that will help you choose an SSD that meets the needs of your applications, makes the most of your available budget and reduces the frequency of future headaches you may have.

An SSD is an SSD, right? Wrong. In fact, SSDs are designed in many different ways to handle very different types of workloads. There are SSDs designed to handle mostly read operations, for example client-class drives, and those that are designed for write-intensive environments. It is extremely important that you understand what applications you are supporting in your data center and the mix of read/write operations required by those applications before you start your search. If you do not have this information on hand when you start your search, the chances that you begin evaluating SSDs that will not be able to deliver what you need will go up significantly, and likely lead to headaches in the future.

Calculating TCO for a solid state drive can be overwhelming, but that’s no reason to shy away from doing so. It is actually extremely important to look at TCO over acquisition costs as the initial price is just that, the initial price. It may be appealing now, however if the drive can’t meet your long-term needs and ends up burning out earlier than anticipated in your environment, you are stuck spending the acquisition cost once again to replace the drive.

The Storage Networking Industry Association (SNIA) has developed tools to help enterprise organizations make TCO calculations, which can be found here. In general though, many organizations are comfortable calculating direct costs like acquisition price, labor and capital expenditures because they are relatively easy to measure. However, you must also take into account indirect costs like lifetime power and cooling, average drive lifespan, replacement costs, maintenance, etc., otherwise you can quickly find yourself in the middle of a TCO nightmare as SSDs burn out after being subjected to application workloads they were not designed to handle.

It’s easy to look immediately at the price of an enterprise SSD and figure out what the best deal is per gigabyte. This is how evaluation of traditional hard drives has always been done. And, when you go to present the options to your boss they will surely love seeing a low $/GB ratio. However, this measurement is not appropriate for SSDs because they have finite lives. They come out of the box with a set number of write operations they can perform before the flash wears out. Because of this, you should look at the cost per Terabyte Write ($/TBW) instead. This will tell you how much you will be paying for the amount of data that you can actually write to the drive over its lifespan. Using this stat for comparison can quickly lead to some amazing realizations, including the fact that the SSD with the lowest acquisition cost can often be the most expensive per TBW. Not exactly a good sign if your applications require a lot of writes.

The power and cooling of your storage setup has a bigger impact than you might think! A metric you should pay attention to is IOPS/Watt. The power density plays a huge part into saving money on energy requirements.  The goal is achieve the highest throughput for the lowest amount of power needed to operate the SSD – especially in tier 0 and 1 systems. Not all SSDs are created equal on this front and since power and cooling efficiency can help save up to 80 percent in total energy expense, this becomes an important metric in organizations with limited resources.

Many IT managers look to SSDs because they offer the performance benefits needed from today’s applications. Often times the latency improvement alone can make all the difference. When looking at latency, it is important to look beyond just the average response time metric. Sure this number should be as low as possible, however you also need to understand how consistent the SSDs latency is. Does the drive experience latency spikes or does latency stay within a relatively tight range? You should always be looking for as consistent of latency as possible, otherwise your data center will perform slower than anticipated while the SSD works through spikes in latency, leaving data sitting, waiting.

Taking all of these items into consideration and doing your homework in advance of your search will not only lead to you selecting an SSD better suited for your environment, but also make it a smoother, more enjoyable search. You will also experience far fewer surprises in the future.

Industry Perspectives is a content channel at Data Center Knowledge highlighting thought leadership in the data center arena. See our guidelines and submission process for information on participating. View previously published Industry Perspectives in our Knowledge Library..