White Paper by John Webster.
The transition from a dependence on rotating disk to solid-state storage within the enterprise data center has dominated the attention of those interested in accelerating storage performance to meet the increasing demands of enterprise applications. And as the cost per unit of solid state storage capacity inevitably decreases, enterprise IT has realized that an investment in solid state storage will drive more revenue generating transactions every business day.
However, improving the performance of the storage media—which the current trend of replacing of disk with flash essentially does—only addresses one aspect of the I/O “stack.”
Another attempt to streamline the I/O path is by eliminating processes designed to make solid state look like disk to operating systems. The coming adoption of the NMVe standard is one example.
A third and perhaps more basic approach to radically improving I/O performance is I/O parallelization. This approach recognizes the shift from single to multicore servers by parallelizing the I/O across cores. Using I/O parallelization to take greater advantage of the processing power offered by today’s multicore servers may not seem on the surface to offer dramatic acceleration of storage performance. A set of SPC benchmarks just released by DataCore indicates otherwise.