Competitive Hill-Climbing Strategies for Replica Placement in a Distributed File System
نویسندگان
چکیده
The Farsite distributed le system stores multiple replicas of les on multiple machines, to provide le access even when some machines are unavailable. Farsite assigns le replicas to machines so as to maximally exploit the di erent degrees of availability of di erent machines, given an allowable replication factor R. We use competitive analysis and simulation to study the performance of three candidate hillclimbing replica placement strategies, MinMax, MinRand, and RandRand, each of which successively exchanges the locations of two le replicas. We show that the MinRand and RandRand strategies are perfectly competitive for R = 2 and 2/3-competitive for R = 3. For general R, MinRand is at least 1/2-competitive and RandRand is at least 10/17-competitive. The MinMax strategy is not competitive. Simulation results show better performance than the theoretic worst-case bounds.
منابع مشابه
Modeling Replica Placement in a Distributed File System: Narrowing the Gap between Analysis and Simulation
We examine the replica placement aspect of a distributed peer-to-peer le system that replicates and stores les on ordinary desktop computers. It has been shown that some desktop machines are available for a greater fraction of time than others, and it is crucial not to place all replicas of any le on machines with low availability. In this paper we study the eÆcacy of three hill-climbing algori...
متن کاملAn Efficient Data Replication Strategy in Large-Scale Data Grid Environments Based on Availability and Popularity
The data grid technology, which uses the scale of the Internet to solve storage limitation for the huge amount of data, has become one of the hot research topics. Recently, data replication strategies have been widely employed in distributed environment to copy frequently accessed data in suitable sites. The primary purposes are shortening distance of file transmission and achieving files from ...
متن کاملReplica Placement in Cloud Storage based on Minimal Blocking Probability
Data replica placement is an important issue in the cloud storage system, which includes two key points: how to determine the replica factor and how to select optimal Datanode to store replica. With the provision of file store service and file access service for large scale users, the Datanode in cloud storage system is regarded as services site. According to the basic principle of queuing theo...
متن کاملImproving Data Grids Performance by Using Modified Dynamic Hierarchical Replication Strategy
Abstract: A Data Grid connects a collection of geographically distributed computational and storage resources that enables users to share data and other resources. Data replication, a technique much discussed by Data Grid researchers in recent years creates multiple copies of file and places them in various locations to shorten file access times. In this paper, a dynamic data replication strate...
متن کامل