NR: A Formal Approach to Traffic Grooming in Optical Networks with General Topologies

该项目的目标是开发用于流量疏导的算法和技术的正式框架，以利用光网络技术的最新进展。这些活动的结果将是一套经过正式验证的流量疏导算法，这些算法可以灵活而高效地应用于各种光网络和成本模型，并将解决问题的离线和在线版本。与最近在SONET/SDH环的上下文中关于业务量疏导技术的工作不同，随着通用多协议标签交换(GMPLS)技术(当交换基于波长时也被称为多协议波长交换，MPLambdaS)的发展，该项目将特别关注具有一般拓扑的光WDM网络。因此，出现了一个新的问题，其中的目标是将一组MPLS业务流整理成光路，以便在一般拓扑的MPLS网络上传输，从而最小化总体网络成本。这一领域还有许多悬而未决的问题有待解决。具体目标包括改进光路由模型以纳入流量疏导，以及评估光网络中降低的波长要求和增加的硬件成本之间的权衡。此外，该项目将在一些成本模型下为基本拓扑制定最优或接近最优的流量疏导策略，还将通过分解为基本拓扑构建块或从基本拓扑构建块合成来研究一般拓扑的疏导。人们普遍预计，这项工作的结果将使未来网络和信息基础设施的高速主干受益，这些网络将由利用混合光学和电子交换实现高性能MPLS传输的网络组成。

The goal of this project is to develop a formal framework of algorithms and techniques for traffic grooming to exploit recent advances in optical network technology. The results from these activities will be a suite of traffic grooming algorithms with formally verified properties, which can be flexibly and efficiently applied within a variety of optical network and cost models, and which will address both the off-line and on-line versions of the problem. In contrast to the recent work on traffic grooming techniques that was in the context of SONET/SDH rings, with the development of generalized multiprotocol label switching (GMPLS) technology (also referred to as multiprotocol lambda switching, MPLambdaS, when switching is based on wavelengths), this project will particularly focus on optical WDM networks with a general topology. Consequently, a new problem arises in which the objective is to groom a set of MPLS traffic streams into lightpaths for transport over an MPLS network of a general topology so as to minimize the overall network cost. There are many unanswered questions yet to resolve in this field. Specific objectives include the refinement of optical routing models to incorporate traffic grooming, and the evaluation of trade-offs between decreased wavelength requirements and increased hardware costs in optical networks. Furthermore, this project will develop optimal or near-optimal traffic grooming strategies for elemental topologies under a number of cost models, and it will also study the grooming in general topologies by decomposition into or synthesis from elemental topology building blocks. It is widely expected that the results of this work will benefit the high-speed backbones of tomorrow's network and information infrastructure will consist of networks utilizing hybrid optical and electronic switching to enable high-performance MPLS transport.