Scalable Architecture for SDN Traffic Classification

Abstract

Scalable network-wide traffic classification (TC), combined with knowledge of endpoint identities, will enable the next wave of innovation in networking, by exposing a valuable layer of TC-enriched flow context information for applications to consume. However, full TC services over standard software-defined networking (SDN) architecture, where packets requiring application services must pass via the control plane, do not scale. Sending packets that require TC inspection to a TC application via the control plane causes resource exhaustion on the control plane (e.g., CPU and bandwidth), and this in turn severely degrades network performance, thus limiting the scalability of such an approach. In this paper, we demonstrate that full TC in SDN can be designed to scale. We describe our scalable TC architecture built atop of widely available OpenFlow v1.3, using a system-based approach. We demonstrate improvements to performance and scalability of SDN TC through experimental results from our prototype implementation. Our intention with this work is to run higher complexity classification functions and demonstrate how this approach benefits TC use cases, as well as SDN standardization.