For most of information technology’s 70-year history since the Eniac computer, the industry has depended on Moore’s Law — the reality that processor chip powery would double roughly every 18 months — to meet the constant demands for higher performance to support new and more complex workloads. Now chip design is reaching its physical limits, and Moore’s Law is breaking down, forcing designers to look elsewhere to satisfy performance improvement demands.

Most of that focus has shifted to storage, writes Wikibon CTO David Floyer. Three things are happening, roughly in parallel:

1. Server SAN, in the form of converged and hyper-converged systems, is replacing networked storage arrays, bringing storage much closer to compute.
2. Flash storage is replacing spinning disk, providing order-of-magnitude faster seek, read and write performance and huge increases in I/O.
3. I/O between flash storage and CPUs is being parallelized to take maximum advantage of the huge bandwidth increase enabled by flash.

These three trends work together to create major improvements in performance. Floyer delineates their advantages and discusses other potential sources of performance improvement on the near horizon. Wikibon premium members can read the full Professional Alert here. To learn about subscribing, look here. Wikibon is a sister company of SiliconANGLE.

Photo credit: kevin dooley via photopin cc