On affinity based routing in multi-system data sharing

Abstract

Multiple systems coupling incurs performance degradation due to inter-system (global) lock contention and database buffer invalidation. At high transaction rates, the level of inter-system interference can have a severe impact on performance. In this paper, we propose a scheme for transaction routing that reduces inter-system interference while keeping load nearly balanced. The routing decision is based on affinity relations defined between transactions and databases. A methodology, employing an integer linear programming technique, is developed to classify incoming transactions into affinity groups based on their database call reference pattern. Based on traces from two of IBM's high volume single system customers, we find that, at high transaction rates, the proposed affinity based routing significantly reduces the lock contention probability and leads to a substantial reduction in transaction response time. Further, the reduction in inter-system data contention, produces a large impact on the performance of optimistic type concurrency control.