Fuzzy logic queue discipline processing over bottleneck link

Abstract

Traffic congestion occurs inevitably in aggregated channel when access networks connect to transport network over a bottleneck data transmission link. The reason of congestion is a lack of bandwidth and continuously customer requirements growth in transmission rate to network content. In this article is investigated influence of data transmission link parameters, as an available bandwidth and packet transmission delay, to queue discipline processing in access router. Fuzzy logic controller (FLC) is considered as an active queue management discipline processing with explicit congestion notification (ECN). Traffic congestion is simulated in a bottleneck link between two routers in Network simulator-2 (NS-2) software. Queue discipline processing was estimated by calculation the quality of service indicators: packets loss, average packet delay and packet delay variation. Statistical analysis of simulation result have given the possibility to define dependence between the quality of service indicators and the data link parameters applicable to considered queue discipline.

Keywords

• Active queue management
• congestion avoidance
• fuzzy logic
• FLC
• network simulator
• NS-2
• TCP/IP