Leveraging SIMD parallelism for accelerating network applications

Abstract

Software packet processing frameworks act as critical components in modern network architecture, as the performance of those have vital effect on the quality of the network services. Motivated by the increasing SIMD capabilities and advancing instructions of recent mainstream CPUs, this paper explores a new parallel processing design and implementation of hash-based data structure and algorithm that are frequently used for building network applications. In particular, we propose effective SIMD optimization techniques for the bloom filter and Open vSwitch megaflow cache. Our design reduces memory access latency via proper prefetching and new data structure design to meet the needs of fast data consuming instructions. Our evaluation shows performance improvements up to 162% in bloom filter and 48% in Open vSwitch compared to the scalar code of each.