A Bursts Contention Avoidance Scheme Based on Streamline Effect Awareness and Limited Intermediate Node Buffering in the Core Network

Abstract

In an Optical Burst Switched (OBS) network, data packets sourced from peripheral networks are assembled into huge sized data bursts. For each assembled data burst, an associated control signal in the form of a burst control packet is (BCP) is generated and scheduled at an offset time ahead of the data burst. The offset timing is to allow for the pre-configuration of required resources at all subsequent intermediate nodes prior to the actual data burst’s arrival. In that way, the data burst will fly-by each node and hence no requirement for temporary buffering at all intermediate nodes. An operational requirement of an OBS network is that it be loss-less as in that way a consistent as well as acceptable quality of service (QoS) for all applications and services it serves as a platform can be guaranteed. Losses in such a network are mainly caused by improper provisioning as well as dimensioning of resources thus leading to contentions among bursts and consequently discarding of some of the contending data bursts. Key to both provisioning as well as proper dimensioning of the available resources in an optimized way is the implementation of effective routing and wavelength (RWA) that will seclude any data losses due to contention occurrences. On the basis of the effects of the streamline effect (SLE), that is, effectively secluding primary contention among flows (streams) in the network, we propose in this paper a limited intermediate buffering that couples with SLE aware prioritized RWA (LIB-PRWA) scheme that combats secondary contention as well. The scheme makes routing decisions such as selection of primary and deflection routes based on current resources states in the candidate paths. A performance comparison of the proposed scheme is carried out and simulation results demonstrate its comparative abilities to effectively reduce losses as well as maintaining both high network resources utilization as well as QoS.