Abstract

Methods for evaluating the performance of transaction processing systems as affected by hardware resource contention are described in this chapter. Hardware resource contention has the primary effect on performance, while data or lock contention has a secondary effect. In the case of blocking-oriented concurrency control methods, such as standard locking, the performance degradation is due to transaction blocking, while in the case of restart-oriented methods transaction response time deteriorates every time a transaction is aborted (we assume that aborted transactions are restarted automatically). The wasted processing due to transaction aborts adversely affects the response time of all transactions. Thus it is important to understand how hardware resource contention affects the performance of transaction processing systems, before considering the effect of data contention.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call