<I style="mso-bidi-font-style: normal">Some algorithms ensures data consistency in distributed systems<?xml:namespace prefix = "o" ns = "urn:schemas-microsoft-com:office:office" /><o:p></o:p></I>

Abstract

Data replication is an optimal and important solution, which is used in distributed applications to meet the need of high availability of data, minimizing bandwidth consumption, increasing fault-tolerance, scalability and security, etc. However, replication technique also poses many complicated challenges, in which, ensuring consistency of data is the major difficulty in developing application software. In this article, many algorithms, which have been developed and applied, are researched and measured basing on some important criteria of distributed systems including the number of replicas, the number of circulating messages in the system and the response time. Moreover, we also recommend an algorithm to solve this problem. The algorithm is based on the sequential consistency model and has the advantages of easy implementation, low bandwidth consumption and fast response time.<?xml:namespace prefix = \"o\" ns = \"urn:schemas-microsoft-com:office:office\" />