Tailor-made Concurrency Control - distributed transactions as a case

Abstract

This paper applies a model for distributed databases and transactions with a distinction between global and local correctness criteria. The global requirements per system are weaker than the local requirements per site. The paper presents an application which suits such a two-level division. The main motivation for our investigation is based on the fact that the commonly used correctness criteria for concurrency control and recovery, serializability and total recoverability, are very strict criteria. The use of more relaxed criteria (allowing more true parallel behaviour and more true partial behaviour) is therefore very appealing - as long as this can be achieved without compromising safety or applicability. The main paradigm in our approach is based on the observation that relatively little knowledge about the databases and transactions can lead to major gains in system throughput. This allows specific systems to have more tailormade correctness criteria. We analyse a specialized type of distributed database, the skeleton-database, and a specialized type of distributed transaction, the wander-transaction. Wander-transactions accessing a skeleton-database allow breaks with both the common serializability criterion and the common total recoverability criterion. Our main emphasis here is on the nonserializability aspect. The primary goal of this work is to designate correctness criteria for controlling local and global parallelism. The secondary goal is to specify priority rules for handling local and global criteria breaks. Wander-transactions accessing a skeleton-database experience dynamic priorities. Our resulting concept, priority serializability, gives increased parallelism without compromising safety.