Towards an AQM Evaluation Testbed with P4 and DPDK

Abstract

Active Queue Management (AQM) addresses the problem arising from using unnecessarily large, unmanaged buffers and thus it aims at improving network and application performance. AQM methods introduce different drop policies to proactively drop packets according to queue states and parameters. Though we are living in an AQM renaissance, comprehensive methodology and framework for performance evaluation under realistic traffic loads are still missing. With the advent of P4, description, validation and evaluation of AQM algorithms in a generic framework have become possible since the different drop policies applied by these methods can be implemented in ingress and/or egress control blocks of a P4 program. In this demo paper, we propose an AQM evaluation framework in which AQM algorithms described in P4 language can be executed and evaluated with a modified version of our DPDK-based P4 compiler called T4P4S in a testbed with realistic traffic mixes. The framework is demonstrated with some selected AQM algorithms in a network with a 5 Gbps bottleneck.