Building a Sensible Storage ROI
The traditional metric with which storage has been bought for decades has been $/GB raw, but that metric is increasingly inadequate to capture the true cost of storage:
How much is usable? Arrays vary widely in their conversion from raw storage to usable storage: Alternative RAID geometries have very different space overheads, as does array meta-data. The lack of thin provisioning can squander capacity, and relatively newer technologies (at least for performance storage) in deduplication and compression can 'give back' storage in allowing you to fit substantially more data on a given amount of raw storage. With typical RAID overhead ranges from 20 to 100% depending on scheme, thinner provisioning yielding 50% savings, and data reduction returning 5-20X, you are much closer to true costs with $/GB usable.
- Performance vs. capacity? Further exacerbating the usable issue, for performance-centric environments, it is common to have to over-provision storage aggressively to get the number of spindles necessary to deliver the performance required, called 'short stroking'. In performance-centric environments, $/GB raw often under-estimates costs by 2 - 3x.
- What about power? The average array costs as much to power/rack over a 3-year lifecycle as it does to acquire. This cost was in the noise when all disk arrays took roughly the same amount of power, but now that flash alternatives which take a fraction of the power exist, power becomes a significant part of the overall TCO of any storage array.
- What about space? With typical storage footprints expanding by 25-50% per year, each years annual purchases can force a costly data center expansion or worse, a new data center or colo. Flash storage is radically smaller as well as more power efficient---think expanding your data footprint 5-10 fold within your existing data center.
- Server over-provisioning? Storage doesn't work alone, it is connected to servers, and often poor storage performance is the reason that servers need to be loaded-up with DRAM and servers scaled horizontally. Each additional server means additional hardware cost, software licensing cost, and management cost. Often the biggest ROI component in moving to flash is the server consolidation savings that is effected, but it is often overlooked in the entire procurement justification.
Given these challenges with the limited $/GB view of storage cost, there's the potential for substantial improvements to the bottom line by weighing broader tradeoffs. The following is a simple model for looking more comprehensively at your storage total cost of ownership (TCO):
At Pure Storage, we have a very simple target: we believe that all-flash storage can be significantly less expensive than traditional spinning-disk storage on a $/GB usable basis, typically 50% the cost of performance disk storage. And if you are willing to also account for space and power savings, as well as for the server consolidation benefits of faster storage, you are likely to find you can save another 50% by moving from spinning disk to all-flash. Even more compellingly, the above analysis assumes only a 5X data reduction, a target Pure Storage has exceeded on nearly all of the customer workloads we have tested to date.