If you read last week's blog about Flash Storage, you may remember I shared some cost comparisons between the IBM FS series and a prominent competitor. This week, I'll cover some smaller items worthy of mention, which also demonstrate more reasons to buy IBM versus the competitor's storage.
Evergreen – This was a GREAT idea by the competition, and it's right to give them credit. However, Evergreen is no longer exclusive to their storage solutions. IBM’s approach to Evergreen is called the “Storage Assurance” option for Flash Systems. I'll admit the name is not as clever, but the reality is that IBM delivers a unique capability to customers who prefer a buying model that renews over time and offers predefined costs over time.
Active/Passive vs Active/Active – IBM supports active/active on all the FS models, whereas the competition offers Active/Passive. Does this matter? Let’s assume both IBM and the top competitors made a processor that could do the exact same amount of work.... Say 10 GB/second sustained, for the sake of argument. The competition's solution with only two processors would deliver 10 GB/second active, with a 10 GB/second standby capability. So, if the primary fails, you still get 10 GB/second of performance. An IBM solution would deliver the same two processors, but in an Active/Active mode. They would deliver 19 GB/second (two x 10 GB/second, less a 5% MP (Multi Processor) effect (for overhead). The same processors in IBM’s design would deliver 19 GB/second instead of 10 GB/second with the competition – for 99.999% of the time!
It is true that if an IBM processor failed, IBM performance would drop to 10 GB/second. Would users notice? Yes. But they would still be getting no less performance than the 10GB/second the competition would deliver. So active/active delivers day in and day out value to your business that the competition does not.
Let’s add to that clustering, specifically the clustering of two, two-way systems. With this configuration, two systems (each with dual processors) can be coupled. So, per the example above, think of two 19 GB systems, also running active/active. Let’s also assume another 5% for MP overhead (which is dramatically overstated for simplicity). That means a cluster of four processers can deliver 36 GB/second using the same technology as in the example above. That’s value. The competition cannot do this.
Ingenuity – FlashCore Modules (FCM). While the competition can provide a few slots within the array which provide slightly higher performance using their Direct Flash Modules (DFM), the array is limited to the number of slots that it has for these flash boards. Also, the DFM’s are proprietary to each array, meaning your expenditure for DFM’s is both limited in scope and will not be able to be used in your next, higher performance array model. What if a standard 2.5” drive form factor could provide the fastest latency, the best performance, the best cost and the ability to move from any FlashSystem to any FlashSystem? This is the FlashCore Module.
Now, what if that same drive could provide the densest enterprise capacity in a 2.5” format and further provide consistent speeds, even when using performance-eating features like compression and encryption, with no loss of throughput? FCM’s are not just dumb drives, but computational storage modules. Modules which increase performance with every drive added.
IBM’s FCM design is itself worth an hour of your time just to understand how it sets IBM flash apart. Ask about it…
Click here for more information on IBM Flash storage and click here for more information on LRS infrastructure solutions.