DMap: A Fault-Tolerant and Scalable Distributed Data Structure

Abstract

Major efforts have been spent in recent years to improve the performance, scalability and reliability of distributed systems. In order to hide the complexity of designing distributed applications, many proposals provide efficient high-level communication abstractions (e.g., atomic multicast). These abstractions, however, are often unfamiliar to average application designers and, as a result, implementing distributed applications that tolerate failures and scale performance without sacrificing consistency remains a challenging task. In this paper, we introduce DMap, a reliable and scalable distributed ordered map. DMap fully implements the generic Java SortedMap interface and can be easily used to scale existing Java applications. To substantiate our claim, we have used DMap to turn H2, a centralized database, into a scalable and reliable data management system.