What is flash storage, and why do I need it?
If you’ve been in this business long enough, you know the Solid State Drive (SSD) has been around since the late 70s/early 80s. Fast and unbelievably reliable compared to spinning disk drives, SSDs for many years were a favorite only at financial institutions, investment houses, and stock markets – some of the few companies that could afford them and needed the speed and reliability of a non-stop system.
Today, organizations of all shapes and sizes can benefit from the ultra-high speeds and friendly price tag of SSDs. But it’s crucial to understand the uses of flash storage in a data center environment and how it relates to your organization’s business needs. That way you can make the right decisions about solid state drives and related storage units.
Speeding up a data center with flash
Flash has come to the forefront of data center storage, and the reason is simple: The need for speed. The data center has changed from a silo-based environment to the modern data center that’s running a plethora of virtualized machines and systems.
As organizations store content on faster servers to consolidate the silos of the data center, the bottlenecking issue becomes storage. Applications and end users need to have their data faster to keep up with the speed at which the servers are processing that data and requests by the applications. Serial Attached SCSI (SAS) drives have maxed out at about 160 I/O operations per second (IOPS) and Near Line SAS (NL-SAS) drives top out at 80 IOPS. In other words, we need more IOPS. We can get there with faster storage, and while 10GbE and 16GbFC connectivity has helped tremendously, we still need more speed. After all, if the data can’t get to the network fast enough, productivity is squandered.
Fortunately, like the superhero, flash comes to the rescue. It takes two forms — cache and data storage. Cache is the extension of the “memory” in the controller. Once the data hits the cache, it remains there as long as it’s “hot” or active. After the data is aged and is no longer “hot,” it gets written or “flushed” to the spinning disk.
SSDs are used as part of the tiered storage environment. Hot data is stored on the SSDs for a certain period of time (often just 24 hours) and then algorithms determine whether it stays in SSD or moves to a lower tier of data.
Flash: Lightning fast speeds with a friendly price tag
Here’s the good stuff. We could spend more time on speeds, feeds, and durability, but there’s one specification to remember: SSDs can do up to 4,000 IOPS per drive, while SAS does about 160. Traditionally, in order to achieve 4,000 IOPS, organizations would need about 25 SAS drives maxed out at 160 IOPS. Even if they didn’t need all the capacity, organizations would need to buy all those drives just to get that performance.
That’s where flash comes into play: Organizations see the same performance in one drive. Here’s a brief example that illustrates the high-performing, cost-saving power of flash storage.
A customer needs 12,000 IOPS for virtual desktop infrastructure (VDI) implementation. There are two ways to get it — SAS and flash (SSD).
The winner is clear: Opting for SSD drives, though more expensive per unit, creates a cost savings of more than $56,000 because of its enhanced power.
What to consider before moving to SSD
One problem that SSD drives present is total capacity: SSDs typically fall in at around 400GB, while SAS and NL-SAS can provide up to 6TB of capacity on a single drive. To combat this storage shortage, organizations can opt for a mixed approach with Hybrid Arrays. By using SSDs to achieve the performance needed (that high IOPS figure), organizations can opt for SAS and NL-SAS drives to meet all capacity requirements.
Still, this solution might not work in all situations or for all organizations. Organizations should be asking their IT teams a number of questions: Can we add SSDs to an existing infrastructure or should we carry through with a forklift upgrade? Would installing a flash array and using the existing storage array solve the problem? Ultimately, is this right for us?
The bottom line is there’s no magic bullet or definitive solution, because every IT environment is different. The needs of the organization, the applications used, and the setup of the environment must be identified and understood before an upgrade occurs.
A third-party supplier with knowledgeable IT professionals can help organizations make that decision. The third party can perform an assessment of the current environment, and help determine if it makes sense to implement SSDs. By examining current infrastructure, outside IT experts can help determine if you should install SSDs to an existing storage area network (SAN), add on a separate flash array, or overhaul with a SAN refresh.
Reach out to your SHI Account Executive for more details, and he or she can engage the resources to get you the answers you need.