IOPS and TB's - cost metrics for each
A while back, I wrote an article that compared/contrasted different ways of measuring storage. In that article, I provided guidance for going beyond “cost per TB” as a primary storage metric and consider that, for I/O-centric environments, cost per IOP might be a better metric. This week, a new product from a startup called IO Turbine is validating that methodology with the release of their Accelio software. The beauty of Accelio’s approach is that it is completely transparent to the guest virtual machines. Accelio works with existing SSD-based storage systems to optimize I/O operations to make sure that the most critical workloads are treated as such and that the overall investment in IOPS as evidenced by the purchase of SSDs is maximized. Although storage capacity is certainly important, overall storage performance is critical, particular as virtual environments are tasked with more and more overall responsibility for supporting the business.