Adaptive Statistical QoS: Learning Parameters to Maximize End-to-End Network Good-put

Author

Evans, Scott C. ; Liu, Ping ; Rothe, Asavari ; Goebel, Kai ; Yan, Weizhong ; Weerakoon, Ishan ; Egan, Marty

Author_Institution

GE Res., Niskayuna, NY

fYear

2006

fDate

23-25 Oct. 2006

Firstpage

1

Lastpage

7

Abstract

We present an adaptive QoS system that seeks to maximize end-to-end success through learning algorithms that take queue depths as input to control weighted fair queue provision. Utilizing an analytical model we generate queue depth and E2E success data for various levels of load and WFQ provision and generate a WFQ provision surface for two classes of real time traffic using neural network techniques. We verify the nature of the surface through event driven simulation and discuss future opportunities for adaptive QoS policy management

Keywords

adaptive systems; learning (artificial intelligence); neural nets; optimisation; quality of service; queueing theory; real-time systems; statistical analysis; telecommunication network management; telecommunication traffic; E2E; WFQ; adaptive statistical QoS; end-to-end network; event driven simulation; learning algorithm; maximization; neural network technique; policy management; quality of service; queue depth; real time traffic; weighted fair queue provision control; Adaptive control; Adaptive systems; Analytical models; Communication system traffic control; Control systems; Neural networks; Programmable control; Queueing analysis; Traffic control; Weight control;

fLanguage

English

Publisher

ieee

Conference_Titel

Military Communications Conference, 2006. MILCOM 2006. IEEE

Conference_Location

Washington, DC

Print_ISBN

1-4244-0617-X

Electronic_ISBN

1-4244-0618-8

Type

conf

DOI

10.1109/MILCOM.2006.302133

Filename

4086778