DESIGN & IMPLEMENTATION OF A PROTOTYPE FOR ANON-REPLICATED DISTRIBUTED DATABASE SYSTEM

Abstract

To study different aspects of distributed database systems a. prototype for such a system has been designed and successfully implemented. The system is capable of supporting a collection, of non-replicated distributed databases spread over a number of computers. It permits fragmentation and location transparency and data in any node can be accessed by submitting queries in a subset of the relational query language SQL. The lexical analyzer and parser for the language have been developed using the utilities Lex and Yacc. The SQL interface, transaction manager and noda.l data manager are the main modules of the system that co-operate to execute a. query. Algorithms 'for converting a.• global, query to subquery messages and integrating the result of these subqueries to execute the global query have been developed and implemented. All c,ammunica.tion between transaction managers and nodal data man'agers is by message passing using message queues and stream socket facilities available in UNIX,. The system has been developed using the C language on HP9000 computer systems