Paper Title: Open Source Data Reduction for High Performance Flash Storage

Paper Abstract: The purpose of data reduction technologies is to drive down the effective cost of storage. The open source dm-vdo project (github.com/dm-vdo) is a good example of a technology explicitly designed to address the high performance flash storage environment. Its Linux kernel-level kvdo module is implemented from the ground up to provide inline, block-level deduplication and compression which the developers have optimized for storage environments with fast random access properties such as today’s NVMe-based devices. There are many challenges related to developing practical data reduction for high performance flash storage environments. The dm-vdo implementation has addressed these requirements through parallelization, queuing, and caching strategies. Users must be able to deploy and tune dm-vdo components to take the best advantage of these strategies in high performance flash storage deployments.

Paper Author: Louis Imershein, Principal Product Manager, Red Hat

Author Bio: Louis Imershein is Principal Product Manager for data reduction technologies at Red Hat and a major contributor to the open source dm-vdo project. At Red Hat, he contributes key requirements and design guidelines that drive development of the Virtual Data Optimizer (VDO) component of Enterprise Linux. Louis has over 30 years of experience in operating system and storage software product management, design, and implementation. He has presented at many conferences around the world, including past Flash Memory Summits. Louis has a Bachelor’s Degree in Biological Science from the University of California, Santa Cruz and has earned certificates in software design, device driver development, system administration, and software testing.