ResearchBib Share Your Research, Maximize Your Social Impacts
Sign for Notice Everyday Sign up >> Login

An Adaptive Leaky-Bucket Mechanism for Traffic Management in OBS Networks

Journal: International Journal of Engineering Sciences & Research Technology (IJESRT) (Vol.3, No. 2)

Publication Date:

Authors : ; ;

Page : 949-955

Keywords : : Leaky Bucket; Routers; Network congestion; Optical Burst Switch; Usage Parameter Control.;

Source : Downloadexternal Find it from : Google Scholarexternal

Abstract

In Optical Burst Switched networks, each light path carry huge amount of traffic, path failures may damage the user application. Hence fault-tolerance becomes an important issue on these networks. Blocking probability is a key index of quality of service in Optical Burst Switched (OBS) network. The Erlang formula has been used extensively in the traffic engineering of optical communication to calculate the blocking probability. A combined preventive/reactive control scheme improves the condition of packet loss due to congestion in networks. The transmission delay and the throttling rate are the major parameters which affect the performance of the reactive control. High throttling rates are most efficient for fast congestion recovery, although sometimes resulting in underutilization of the link. A combined reactive/preventive congestion control mechanism is investigated in this paper with emphasis on the Leaky Bucket (LB) mechanism chosen for source traffic policing in computer networks. The fluid-flow model is used to analyze the performance of both buffered and un-buffered LBs. It is proposed that one LB is not sufficient to manage all the source traffic parameters. If tight control, fast reaction time and a small queuing delay are required then according to the analysis done, the proposed triple LB mechanism is an effective solution. According to the delayed congestion feedback information received from the network the LB parameters are dynamically changed. The preventive control policy is compared with the adaptive control scheme. The results show that even for large propagation delays, major performance improvements are possible by using an appropriate feedback policy.

Last modified: 2014-09-03 22:33:44