Distributed Concurrency Control Analysis for Data Sharing

Abstract

The performance of multiple systems sharing a common data base with distributed locking is analyzed. The analysis is based on traces from large mainframe systems running IBM's IMS database management system. A methodology to synthesize a multi-system IMS lock trace from single-system traces is presented. The multi-system traces are used in trace driven simulations to predict lock contention in a multi-system environment and to generate workload parameters. These parameters are used in event-driven simulation models to examine the overall performance of two-system and four-system data sharing, and to determine the performance impact of various system parameters and design alternatives. While performance results are presented for realistic system parameters, the emphasis is on the methodology, analysis technique and on examining the factors that affect multi-system performance.