Adaptive statistical sampling of VOIP traffic in WLAN and wired networks using Fuzzy Inference System

Abstract

Real-time audio and video applications generate vast amount of data in form of information packets. The collection and processing of all these packets in real time are not practically feasible. Therefore, an appropriate sampling technique is required in order to reduce the amount of collected data and their processing. In this paper, a statistical adaptive sampling technique to adjust sampling rate based on the traffic's statistics was developed using a Fuzzy Inference System (FIS). The FIS determined the sampling rate by using a set of rules to interpret statistical variations in Quality of Service (QoS) parameters. A comparison of adaptive statistical sampling using FIS against systematic, stratified, and random sampling was also carried out. The Network Simulator- 2 (NS-2) was used to evaluate the operation of sampling techniques. The study indicated that biasness values of sampled traffic obtained from the developed adaptive sampling technique were closer to zero than the values obtained using conventional sampling techniques which indicates the effectiveness of the proposed method.