Workload characterization of stateful networking applications

Abstract

The explosive and robust growth of the Internet owes a lot to the "end-to-end principle", which pushes stateful operations to the end-points. The Internet grow both in traffic volume, and in the richness of the applications it supports. A whole new class of applications requires stateful processing. This paper presents the first workload characterization of stateful networking applications. The analysis emphasizes the study of data cache behaviour. Nevertheless, we also discuss other issues, such as branch prediction, instruction distribution and ILP. Another important contribution is the study of the state categories of the networking applications. The results show an important memory bottleneck that involves new challenges to overcome. © Springer-Verlag Berlin Heidelberg 2008.