Top 5 Benefits of All Flash Storage
Since flash storage was introduced into the mainstream SAN marketplace less than a decade ago, its presence in data centers across the world has grown like a weed and shows little signs of slowing down. With All-Flash SANs becoming more prevalent but still not the norm, we thought we’d go over the current top 5 benefits of all flash storage.
Perhaps the most well-known benefit of moving to an All-Flash solution is the latency, or lack thereof. All-Flash solutions typically see sub 1ms latencies on a mixed workload, and seeing peak latency above 5ms is extremely rare. Compare this with hybrid SANs which see an average latency around 5-10ms, or an all spinning disk solution which has a very wide range of latency possibilities, and the advantage is clear. With latencies this low, users experience a performance like no other. This allows them to access the applications and data they require at unparalleled speeds, and as a result, their productivity and efficiency can sky rocket!
02. Virtual Desktops
VDI is a growing trend for businesses of all sizes, but all too often businesses have had to say “no” to VDI because of the performance uncertainty. VDI requires a lot of IO, and the more users an organization has, the more IOPS that will bog down the system. In addition to a normal workload, VDI users add a lot of performance overhead to the centralized SAN. For example, most organizations are comprised of “knowledge users,” these are employees that use applications like Microsoft Office, internet, CRM, video streaming and some others. Knowledge users produce about 25 IOPS per desktop on average, so multiply the number of users by 25, and the amount of extra IO on the SAN adds up quickly! An organization of 100 users will produce an additional 2,500 IOPS during working hours. An organization of 500 users will add 12,500 IOPS, and an organization of 1,000 users will add 25,000 IOPS. These numbers grow so much, so quickly that many organizations who utilize VDI have to do so on a dedicated SAN, separate from their centralized node. All-Flash allows for an organization to host all of their applications on a single device, eliminating unnecessary network complexity.
This one seems a little odd, right? Since SSDs are typically offered in smaller capacities than their HDD counterparts, it feels weird to say that capacity is a benefit of all flash. Well the truth is that SSDs have grown in capacity a lot in the last few years. Multiple vendors now offer 4TB SSDs and the price of SSDs continues to drop (with the exception of that unique shortage of SSDs at the beginning of 2017). As solid state technology continues to advance and the prices continue to drop, the capacity benefits and value will only continue to increase!
One of the most overlooked benefits of All-Flash storage, the reliability of SSD based storage, is one of its best features. SSDs don’t fail in the traditional sense like HDDs do. Instead, SSDs run out of endurance and eventually “burn out.” Luckily for us, they don’t do this often because most vendors account for their drive endurance with sophisticated algorithms that make sure no one SSD is written to more often than it can handle.
Finally, failures, whether physical or software related, can cause havoc upon an infrastructure. Depending on the storage array and type, a disk failure can be a simple hot swap component, or in some cases, can slow a storage array and come one step closer to data loss. If a double or triple drive failure occurs, hopefully things like hot spares are available and the RAID can rebuild itself automatically. Otherwise, that backup solution mentioned earlier will come in quite handy. Other things like power supply, NIC, CPU, memory, or complete controller failure can cause a storage array to go down, or in some cases become overloaded when failover occurs. If the system is running in an overloaded state, thresholds will be hit and latency will increase across the infrastructure. IT Administrators need to be prepared so that if a failure occurs, there is a plan in place to recover, failover, or at worst case revert back to a stable point.The storage solution in place must have sufficient management tools to reduce time and complexity of typical tasks, built in reporting and threshold monitors to alert in the event of impeding catastrophes, integration with backup or the ability to replicate in the event of site loss, and of course meet the budgetary requirements within the scope of the plan..