A performance study of divergence control algorithms

Epsilon serializability (ESR) was proposed to relax serializability constra ints by allowing transactions to execute with a limited amount of inconsist ency (epsilon-spec). Divergence control algorithms, viewed as extensions of concurrency control algorithms, enable read-only transactions to complete if their inconsistencies do not exceed epsilon-spec. This paper studies the performance of two-phase locking divergence control (2PLDC) and optimistic divergence control (ODC) algorithms. We develop a central part of the ESR transaction processing system that runs with 2PLDC and ODC. We applied a co mprehensive centralized database simulation model to measure the performanc e. Evaluations are conducted with multi-class workloads where on-line updat e transactions and long-duration queries progress under various epsilon-spe c. Our results demonstrate that significant performance enhancements are ac hieved with a non-zero tolerable inconsistency. With sufficient epsilon-spe c and limited system resources, both algorithms result in comparable perfor mance. However, with low resource contention, ODC performs significantly be tter than 2PLDC. Furthermore, in the range of small epsilon-spec, the queri es committed by ODC have more accurate results than those committed by 2PLD C.