Concurrent and Reliable Updates of Distributed Databases

Abstract

A concurrent execution of transactions and various failures occuring during transaction processing in a distributed database system can lead to an inconsistent database state. In order to prevent such inconsistency from occuring , 1) the schedule of transactions must be equivalent to some serial schedule and 2) each transaction must be either completed or backed out. This paper develops a set of schemes that satisfy these requirements and still realize highly concurrent execution of transactions. This paper also shows how to incorporate these schemes into a multi-level distributed database system where there exists a hierarchy of transactions. Detailed algorithms for concurrent and reliable updates of distrubuted databases based on the proposed schemes are included in the appendix.