Achieving Fairness in Generalized Processor Sharing for Network Switches

In Generalized Processor Sharing (GPS), the goal of the scheduler is to maximize the processor utilization, while maintaining a rate guarantee and fairness in the bandwidth allocation to each input stream according to the quality of service (QoS) criteria. While in the single server case, these requirements can be satisfied simultaneously by scheduling service rate according to a scale version of the rate request, the result does not generalize to the case when multiple processors are used. This paper extends the notion of max-min and proportional fairness for single node Generalized Processor Sharing in Network Switches (GPS-SW), and provides rate augmentation algorithms in achieving the two performance criteria. Keywords— Switching and Scheduling, Generalized Processor Sharing, Max-Min Fairness, Proportional Fairness

[1]  Frank Kelly,et al.  Rate control for communication networks: shadow prices, proportional fairness and stability , 1998, J. Oper. Res. Soc..

[2]  Sanjoy K. Baruah,et al.  The case for fair multiprocessor scheduling , 2003, Proceedings International Parallel and Distributed Processing Symposium.

[3]  Cheng-Shang Chang,et al.  Birkhoff-von Neumann input buffered crossbar switches , 2000, Proceedings IEEE INFOCOM 2000. Conference on Computer Communications. Nineteenth Annual Joint Conference of the IEEE Computer and Communications Societies (Cat. No.00CH37064).

[4]  Dimitri P. Bertsekas,et al.  Data Networks , 1986 .

[5]  Analysis and Simulation of a Fair Queuing Algorithm , 2008 .

[6]  Abhay Parekh,et al.  A generalized processor sharing approach to flow control in integrated services networks: the single-node case , 1993, TNET.

[7]  Banu Ozden,et al.  Fair queuing for aggregated multiple links , 2001, SIGCOMM 2001.

[8]  John N. Tsitsiklis,et al.  Introduction to linear optimization , 1997, Athena scientific optimization and computation series.

[9]  Harish Sethu,et al.  On achieving fairness in the joint allocation of processing and bandwidth resources , 2003, IWQoS'03.