Performance of generative programming based protocol implementation

摘要

The protocol implementation framework for Linux (PIX) is a protocol development tool using generative programming. It aims at capturing the similarities in behavior among different layers of protocols and grouping solutions to cross-cutting concerns of communication systems. It achieves a high degree of configurability by providing several combinations which could be chosen to generate the desired protocols. The paper investigates how the performance of the generative programming based protocol implementation compares with traditional protocol implementation techniques. A benchmark is developed to give a thorough performance analysis of PIX to contrast it with other protocol development frameworks. The benchmark compares the performance of bulk data transfer. The file transfer protocol (FTP) is used for comparison purposes. Latency, throughput and resource usage measurements are provided in order to compare the performance of PIX and generative programming with NcFTP, which uses structured programming, and x-Kernel, which uses structured and object-based programming.