In this paper authors propose a fair queuing (FQ) algorithm to mainly provide 1) fair allocation of bandwidth, 2) lower delay for sources using less than their full share of bandwidth, and 3)protection from ill-behaved sources. They argue that First-Come-First-Serve (FCFS) queuing is not adequate; more discriminating queuing algorithms must be used in conjunction with source flow control algorithms to control congestion effectively in noncooperative environments. While designing algorithm they consider mainly three things 1) bandwidth, 2) promptness, and 3) buffer space. They conduct tests using six different scenarios comparing FQ against FCFS using generic, Jacobson and Karels (JK), and DECbit flow control algorithms. Test results show that FQ provides reasonably satisfactory congestion control for given scenarios.
There are several contributions of this paper. First of all, FQ algorithm provides low delay to sources which use less than their fair share of bandwidth. Furthermore, their algorithm also targets the problems caused by ill-behaved sources. With FCFS, ill-behaved sources may cause serious problems causing other flows almost shut out. However, with FQ they obtain their fair share of bandwidth. This property also makes FQ gateways to be considered as a kind of firewalls that protect users and rest of the network from ill-behaved sources.
Although this initial FQ algorithm provides a mean fairness to users, it may not provide complete fairness in today’s complex network conditions. This is because initial FQ algorithm only tries to achieve fair allocation of bandwidth but there are several other things shared by users/flows such as links, buffers, CPUs etc. Therefore, in order to provide overall fairness other factors also should be considered. Also this algorithm may not perform well for wireless system where underlying output link conditions may vary frequently.
Please join StudyMode to read the full document