A Simple Approach to Implementing Computational Storage | SNIA

Abstract

Ever-increasing storage workloads require a shift to computational storage solutions that bring compute closer to the stored data. Data-driven applications that benefit from database searches, data manipulation, and machine learning can perform better and be more scalable if designers add computation directly to storage. An easy way to make this distribution work is to simply load the relevant data onto individual solid-state drives and take advantage of the increased compute capabilities of low-cost local microcontrollers. Several methods can be used to take advantage of computational storage by provisioning a drive and accessing it from the host. Designers can use host commands or open-source software-based solutions.