A wait-free algorithm for optimistic programming: HOPE realized

摘要

Optimism is a powerful technique for avoiding latency by increasing concurrency. Optimistically assuming the results of one computation allows other computations to execute in parallel, even when they depend on the assumed result. Optimistic techniques can particularly benefit distributed systems because of the critical impact of communications latency. This paper reviews HOPE: our model of optimistic programming, and describes how optimism can enhance distributed program performance by avoiding remote communications delay. We then present the wait-free algorithm used to implement HOPE in a distributed environment.