SWM: Simplified Wu-Manber for GPU-based Deep Packet Inspection

Graphics processing units (GPU) have potential to speed up deep packet inspection (DPI) by processing many packets in parallel. However, popular methods of DPI such as deterministic finite automata are limited because they are single stride. Alternatively, the complexity of multiple stride methods is not appropriate for the SIMD operation of a GPU. In this work we present SWM, a simplified, multiple stride, Wu-Manber like algorithm for GPU-based deep packet inspection. SWM uses a novel method to group patterns such that the shift tables are simplified and therefore appropriate for SIMD operation. This novel grouping of patterns has many benefits including eliminating the need for hashing, allowing processing on nonfixed pattern lengths, eliminating sequential pattern comparison and allowing shift tables to fit into the small on-chip memories of GPU stream cores. We show that SWM achieves 2 Gb/s deep packet inspection even on a single GPU with only 32 stream cores. We expect that this will increase proportionally with additional stream cores which number in the hundreds to thousands on higher end GPUs.