GPU and other SIMD stream architecture have been used for accelerating packet processing applications. This paper explores the design space on GPU for sketch-based network traffic change detection application by using OpenCL parallel programming framework. Due to the parallel nature of sketch data structure, the computations can be mapped to OpenCL execution model on GPU efficiently. The sketch data structure is mapped to buffer object in device's global memory, and work-items are executed on the sketch in parallel. Compared to the sequential CPU implementation, the experiment results on Radeon HD 5870 GPU show that the hash computation and ESTIMATE operation achieved about 15 times and 9 times speedup, respectively.
Paper available at IEEE.