Adaptive Data Replication in Distributed Systems

伊利诺伊大学芝加哥分校和巴尔的摩县的马里兰州大学在这个项目上合作，其动机是互联网和万维网（WWW）的巨大增长。自适应复制是一种在分布式数据库系统，特别是Internet/WWW中提高性能、响应时间和访问成本的方法。自适应复制是缓存原理的推广。在自适应复制中，对象的副本数量、这些副本的位置以及哪些更新被传播到每个副本，根据给定的成本函数动态地改变。该项目建立了分布式数据库系统中自适应复制的基本原则，并开发、实现和分析了一个对象自适应复制系统。该系统由一组用于各种成本函数的算法组成。为了进行分析，研究人员正在构建一个模拟测试平台，用户在其中定义网络，选择复制算法，选择成本模型以及每台计算机的访问模式。该系统能够在各种自适应和静态复制算法之间进行比较。该项目将提高分布式系统和分布式数据库的理解、成本和性能，并可能导致更有效的互联网和万维网。

The University of Illinois at Chicago and the University of Maryland at Baltimore County collaborate on this project motivated by the tremendous growth of the Internet and the World-Wide-Web (WWW). Adaptive replication is an approach for improving performance, response time, and access cost in distributed database systems in general, and Internet/WWW in particular. Adaptive replication is a generalization of the principle of caching. In adaptive replication the number of copies of an object, the location of these copies, and which updates are propagated to each copy, change dynamically depending on a given cost function. The project establishes the fundamental principles of adaptive replication in distributed database systems, and it develops, implements and analyzes a system for adaptive replication of objects. The system consists of a set of algorithms to be used for various cost functions. For the analysis the investigators are building a simulation testbed in which the user defines the network, selects a replication algorithm, selects a cost model, and an access pattern at each computer. The system enables comparison among the various adaptive and static replication algorithms. The project will improve the understanding, cost, and performance of distributed systems and distributed databases, and it can lead to a more efficient Internet and World-Wide-Web.