Using Replication for High Performance Database Support in Distributed Real-Time Systems

Abstract

Considerable research effort has been concentrated to the problem of developing techniques for achieving high availability of critical data in distributed real-time systems. One approach is to use replication. Replicated data is stored redundantly at multiple sites so that it can be used even if some of the copies are not available due to failures. This paper presents an algorithm for maintaining consistency and improving the performance of database with replicated data in distributed real - t:ime systems. The semantic information of read~only transactions is exploited for improved efficiency, and a rnultiversion technique is used to increase the degree of concurrency. Related issues including the consistency of the states seen by transactions, version management, and recovery of replicated data in distributed systems are discussed. Note: Abstract extracted from PDF file via OCR