Tuple-based abstract data types: full parallelism

TL;DR

This paper introduces a restricted form of commutativity for abstract data types that enhances parallelism and simplifies concurrency control by enabling compile-time determination of operation compatibility.

Contribution

It proposes a novel approach to concurrency control that increases parallelism without additional runtime overhead, avoiding the need for specifying commutativity relations or inverse operations.

Findings

Compile-time commutativity determination

Efficient runtime checking comparable to compatibility

Reduced log space utilization

Abstract

Commutativity has the same inherent limitations as compatibility. Then, it is worth conceiving simple concurrency control techniques. We propose a restricted form of commutativity which increases parallelism without incurring a higher overhead than compatibility. Advantages of our proposition are: (1) commutativity of operations is determined at compile-time, (2) run-time checking is as efficient as for compatibility, (3) neither commutativity relations, (4) nor inverse operations, need to be specified, and (5) log space utilization is reduced.