Competition For Resources In The Obs Network Problems And Solutions

In 1990s network bandwidth got a huge leverage with the developments of WDM and DWDM technologies. But until now switching technologies such as optical circuit switching have to be confined in the electronic domain, and therefore switching nodes have become the bottleneck limiting network speeds.Optical burst switching (OBS) was proposed by Chunming Qiao and J. S. Turner which combined* advantages of optical circuit switching and optical packet switching, it is a balance technology between the two methods. Currently it is regarded as a transition scheme from optical circuit switching to optical packet switching. This paper will concentrate on the research of channel scheduling and burst contention resolution in OBS networks.The first part is an introduction to OCS, OBS and OPS. Their pros and cons are analysed and compared in detail.The second part describes the basic idea, structure and key technologies of OBS, including signaling protocols, assembling technologies, channel scheduling and burst contention resolution strategies.The third part discusses data channel scheduling algorithm in OBS networks. At first two important channel scheduling algorithms commonly used in OBS (LAUC and LAUC-VF) are analysed. Some simulation proves the LAUC-VF has a better performance. Then we investigate priority-based scheduling algorithms. In order to reduce the high rate of invalid resources in the priority mechanism, a RRIT strategy is proposed. Simulation shows that this scheme could significantly reduce packet loss rate and improve bandwidth utilization.In the forth part we study the contention revolution schemes in depth. Based on various contention revolution schemes in OBS, we will propose a very concrete scheme including both non-preemptive and preemptive methods, combining with RRIT strategies, to minimize the so called "false contention" effect. Simulation results show that this new strategy can effectively improve OBS system performance.The fifth part concludes this paper with a summary of works done and to be done.