High-Performance Publish-Subscribe Matching Using Parallel Hardware

Matching incoming event notifications against received subscriptions are a fundamental part of every publish-subscribe infrastructure. In the case of content-based systems this is a fairly complex and time consuming task, whose performance impacts that of the entire system. In the past, several algorithms have been proposed for efficient content-based event matching. While they differ in most aspects, they have in common the fact of being conceived to run on conventional, sequential hardware. On the other hand, parallel hardware is becoming available off-the-shelf: the number of cores inside CPUs is constantly increasing, and CUDA makes it possible to access the power of GPU hardware for general purpose computing. In this paper, we describe a new publish-subscribe content-based matching algorithm designed to run efficiently both on multicore CPUs and CUDA GPUs. A detailed comparison with two state-of-the-art sequential matching algorithms demonstrates how the use of parallel hardware can bring impressive speedups in content-based matching. At the same time, our analysis identifies the characteristic aspects of multicore and CUDA programming that mostly impact performance.