L4S: Low-Speed Software Synergetic Sampling and Detecting Long Flow for Data Center Network

Sampling and detecting long flow are important for traffic engineering in data center network. However, the result of accurate sampling based on hardware is usually associated with high sampling rate. In this paper, we proposed L4 S: a low-speed software synergetic sampling with a self-adaptive sampling rate adjustment algorithm for the problem that the sampling rate cannot match the network traffic and verified the feasibility of L4S in the long flow detection problem, The simulation results show that the L4S reduces the sampling rate from 1000:1 to 10000:1 when it achieves a comparable sampling effect with a relatively independent hardware sampling and a fixed sampling rate scheme.