Abstract

AbstractAn all-OPS network architecture displays a very attractive prospect for the future highly flexible optical transport networks. However, due to a variety of technical challenges, there is still a long way to go to accomplish a mature realization of an OPS network that is ready for practical deployments. Optoelectronic packet switching alleviates the implementation difficult to some degree, but a series of technical innovations is still needed in optical signal processing concerning timing and synchronization. In general, OPS is regarded as the long-term solution for high speed optical networks but, in the meanwhile, more feasible network architectures are desired for the efficient transport of highly dynamic and bursty data traffic. This is the case of optical burst switching (OBS).In OBS networks, a special data unit called burst is defined for the transmission of user data. The data burst is generated by aggregating client traffic at the ingress nodes and it may have a variable size. In the network core, such data bursts are switched and transported asynchronously over the optical domain without any O/E/O conversion. The burst switching/routing decision is performed by the electronic processing of an out-of-band signaling message that arrives prior to its associated data burst. Similar to the optoelectronic packet switching, OBS ensures a transparent switching of the optical data traffic and keeps the signal processing and switching control in the electronic domain. In addition, data bursts are transmitted asynchronously in OBS, thus removing synchronization complexity. Also, design that adopts an offset time between the transmission of the signaling message and its associated data burst makes optical buffering not mandatory in the core nodes, thus improving the feasibility of OBS.This chapter summarizes the fundamentals of optical burst-switched networks and further covers a wide range of the latest research and development issues in the field. Section 4.1 reviews the network and node architecture of OBS networks. Section 4.2 outlines algorithms for burst assembly at the edge node, which have a significant impact on network performance. Signaling and channel reservation protocols in the core network are discussed through Section 4.3. Section 4.4 explores different solutions for resolving contention between data bursts. Section 4.5 investigates the issues of QoS provisioning in OBS networks. Section 4.6 studies the impact of an OBS underlying architecture in the upper layer protocol: TCP. Finally, Section 4.7 summarizes the Erlang Fixed Point (EFP) iterative algorithm that is widely used for the performance evaluation of OBS networks.KeywordsTransmission Control ProtocolPacket Loss RateOptical Burst SwitchingCongestion WindowData BurstThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call