Multilayer Packet Classification With Graphics Processing Units

doi:10.1109/TNET.2015.2491265

Multilayer Packet Classification With Graphics Processing Units

2016-10-01

50

OA

AI

DOI：10.1109/TNET.2015.2491265

Original

Original request for help

Save

Abstract

En 中文

The rapid growth of server virtualization has ignited a wide adoption of software-based virtual switches, with significant interest in speeding up their performance. In a similar trend, software-defined networking (SDN), with its strong reliance on rule-based flow classification, has also created renewed interest in multidimensional packet classification. However, despite these recent advances, the performance of current software-based packet classifiers is still limited, mostly by the low parallelism of general-purpose CPUs. In this paper, we explore how to accelerate packet classification using the high parallelism and latency-hiding capabilities of graphic processing units (GPUs). We implement GPU-accelerated versions for both linear and tuple search, currently deployed in virtual switches, and also introduce a novel algorithm called Bloom search. These algorithms are integrated with highspeed packet I/O to build GSwitch, a GPU-accelerated software switch, and also to extend Open vSwitch. Our experimental evaluation indicates that, under realistic rule sets, GSwitch is at least 7 faster than an equally-priced CPU classifier. We also show that our GPU-accelerated Open vSwitch outperforms the classic Open vSwitch implementation by a factor of 10, on average.

Keywords:

Packet classification

software switch

OpenFlow

software-defined networking

GPU

CUDA

Open vSwitch

Multilayer Packet Classification With Graphics Processing Units

Abstract

Journal

Organization