Availability-aware design for generalized MPLS networks

互联网服务订户，包括企业和普通终端用户，都希望从他们的服务提供商那里获得高可用性的互联网连接。在动态通用多协议标签交换(GMPLS)网络中，保证每个连接的端到端(E2E)可用性一直是一个吸引人的设计目标。以往在相关领域的研究工作主要集中在针对给定保护类型下的单故障或双故障的可生存路由方案的研究。然而，随着最先进的通信运营商网络向多业务带宽分配环境发展，每个LSP的E2E可用性已经成为实现最终用户所接受的服务质量(Qos)保证的主要性能指标。这一事实对可存活路由和备用容量分配方案的开发产生了根本影响，因为终端用户可能不关心运营商可以处理多少同时故障，以及运营商采用什么类型的保护方案。在这些观察的推动下，该项目致力于在GMPLS网络中为实时和非实时服务开发一个通用的可用性感知服务供应架构，其中将定义和操纵每个连接的长期和短期E2E可用性。在可用性模型的基础上，对已报道的生存性路由和空闲容量分配方案进行了相应的改进。这项拟议的研究旨在通过将以前的一些设计问题(如运营商可以处理多少并发故障，以及我们最好采用哪种类型的保护方案)不再作为问题，为可生存路由问题提供完整的解决方案。这种可生存性的路由体系结构有望更好地适应现代通信运营商网络的设计要求，其中服务可用性是冗余投资和生存性保证的最终目标。该项目还将导致HQP在可用性分析和优化技术方面的培训，这将极大地促进加拿大通信业的发展。

The Internet service subscribers, including both enterprises and general end users, expect for high availability of Internet connections from their service providers. Guarantee of end-to-end (E2E) availability for each connection in dynamic Generalized Multi-Protocol Label Switched (GMPLS) networks has been an attractive design objective. The previous research work in the related areas has focused on development of survivable routing schemes for single or dual simultaneous failures under a given type of protection. However, as the state-of-the-art communication carrier networks evolve to a multi-service bandwidth provisioning environment, the E2E availability of each LSP has become a major performance index for achieving Quality-of-Service (QoS) guarantee percieved by the end users. This fact has imposed a fundamental impact on the development of survivable routing and spare capacity allocation schemes since the end users may not care about how many simultaneous failures the carrier can handle, and what type of protection scheme the carrier adopts. Driven by the observations, this project is committed to work on a general availability-aware service provisioning architecture in GMPLS networks for both real-time and non real-time services, where the long-term and short-term E2E availability of each connection will be defined and manipulated. Based on the availability models, the corresponding enhancement on the reported survivable routing and spare capacity allocation schemes will be investigated. The proposed research is envisioned to provide a complete solution to survivable routing problems by leaving some of the previous design issues, such as "how many simultaneous failures the carrier can handle", and "what type of protection scheme we had better adopt", no longer as problems. Such survivable routing architecture is expected to better fit into the design requirements of modern communication carrier networks where service availability is the ultimate goal of redundancy investment and survivability guarantee. The project will also lead to HQP training in availability analysis and optimization techniques, which will greatly contribute to the Canadian communication industry.