This chapter deals with control of network traffic in routers as well as end-to-end flows. First it is proposed an scheme for implementing end-to-end traffic control mechanisms through fuzzy inference systems. A comparative evaluation of simulation and implementation results from the fuzzy rate controler as compared to that of traditional TCP flow and rate control mechanisms is performed for a wide set of realistic scenarios. Then, fuzzy inference systems for traffic control in routers are designed. A particular proposal has been evaluated in realistic scenarios and is shown to be robust. The proposal is compared against the random early detection (RED) scheme. It is experimentally shown that fuzzy systems can provide better performance and better adaptation to different requirements with mechanisms that are easy to modify using linguistic knowledge.

Peer Reviewed