Paper Title: Computational Storage Is the Answer for Huge Data and Deep Problems

Paper Abstract: Computational storage accelerates applications by adding compute power where the data is stored. The results are fewer large data moves, less burden on CPUs and connections, and much greater scalability. Initial trials with large, complex systems programs show remarkable results. For example, by applying computational storage in an NVMe-based system, the RocksDB production in-memory database yields 6x the throughput with half the CPU requirements. Similarly, the widely used ZFS production filesystem becomes twenty times more power-efficient. Like results can be achieved in data-driven applications in AI/ML, HPC, real-time analytics, security, IoT, and content delivery networks. Furthermore, the advantages will surely increase as data stores keep increasing rapidly in size and customers want to delve deeper into problems to gain a competitive advantage.

Paper Author: Andrew Maier, Software Engineer, Eidetic Communications

Author Bio: Andrew Maier has been involved in the NoLoad™ project as a Software Engineer with Eideticom since January 2017. He has a B.Sc. and M.Sc. in Computer Engineering from the University of Alberta and his recent research included the acceleration of LDPC codes using OpenCL for FPGAs. He participated in the international robotics competition BattleBots as part of team MBSRobotics that aired on ABC in the summer of 2016.