Client's cache updating using idle time of clients in relational database management systems

Abstract

Although multi-tiered software development is taking the place in database application environments client-server infrastructure has still been in use for years. Furthermore both infrastructures still need to be improved in some aspects. One of the important aspects of improvement is caching. This paper addresses the issue of caching in relational database system within client-server structure and a method that can be used to keep clients' cache up to date. The main objective of the proposed caching methods is to use clients' idle time in order that clients' cache is up-to-dated and to use the clients' memories for caching. For this purpose a new combined system which includes CCDT (Clients Cache Description Table) for organizing client cache updating predicate-based caching method MNTC (Maximum Number of Tuples in Cache) that includes a managerial patch in server-side and small patches in clients-side and TPCF (Time Period calculated from client Commit Frequency) for calculation of idle time of the clients is presented.