Sample DSpace Document Title

Abstract

Energy has become an increasingly large financial and scaling burden for computing. With the increasing demand for and scale of Data-Intensive Scalable Computing (DISC), the costs of running large data centers are becoming dominated by power and cooling. In this thesis we propose to help reduce the energy consumed by large-scale computing by using a FAWN: A Fast Array of Wimpy Nodes. FAWN is an approach to building datacenters using low-cost, low-power hardware devices that are individually optimized for energy efficiency (performance/watt) rather than raw performance alone. FAWN nodes are individually resource-constrained, motivating the development of distributed systems software with efficient processing, low memory consumption, and careful use of flash storage. In this proposal, we investigate the applicability of FAWN to data-intensive workloads. First, we present FAWN-KV: a deep study into building a distributed key-value storage system on a FAWN prototype. We then present a broader classification and workload analysis showing when FAWN can be more energy-efficient, and under what conditions that wimpy nodes perform poorly. Based on our experiences building software for FAWN, we finish by presenting Storage Click: a software architecture for providing efficient processing of remote, small storage objects.