Design and operation of highly reliable telecommunication networks

诸如交换节点的单个电信网络组件的物理故障，或者甚至更可能的光纤电缆的物理故障可能是灾难性的，影响大量的商业操作，甚至影响紧急系统(911)。即使在正常情况下，这类故障在大型网络中也出人意料地频繁，最近对恐怖主义和蓄意攻击或破坏的担忧只会加剧确保网络可靠性和可用性的方法的重要性。目前在网络生存性问题上所做的许多工作通常假设，尽管故障不一定可以预测，但潜在的故障场景是已知的并且可以预先指定(例如，单个光缆，或者在某些情况下，一次不超过两到三个光缆的组合)。然而，如何确定如何在任意广泛的灾难性故障情况下对网络进行优化设计和动态重新配置，以最大限度地提高生存能力，特别是关键服务的生存能力，仍然是一个非常悬而未决的问题，也是我提议进行的研究的主题。这项研究将扩展现有的生存性原则，在可能的情况下允许灾难恢复，并开发新的方法来处理传统协议无法适应的灾难恢复方面。更重要的是，我们将寻求开发和演示一套生存性机制，这些机制能够从任意广泛的网络故障中恢复，并能够提供协调的网络范围重新配置响应，以最大限度地提高关键服务的生存性，而额外的资本投资最少。分布式自组织路径恢复特别适合于这里所需的全局重新配置类型，并且在这方面，几种新兴的网络生存性方法(包括节点的跨度恢复、故障无关的路径保护P循环、节点环绕的P循环)也显示出希望。这项研究的重点将包括修改现有的协议和开发新的协议，以便能够在全网络范围内对灾难性故障做出协调一致的反应，特别是使关键或紧急通信能够最大限度地恢复。

Physical failure of a single telecommunication network component such as a switching node, or even more likely of a fibre cable, can be disastrous, affecting a vast array of business operations and even emergency systems (911). Such failures are surprisingly frequent in large networks even in normal times, and recent concerns about terrorism and deliberate attack or sabotage only heighten the importance of methods to ensure network reliability and availability. Much of the current work done on the network survivability problem generally assumes that, although failures can not necessarily be predicted, potential failure scenarios are known and specifiable in advance (say, individual fibre cables, or in some cases, combinations of no more than two or three at a time). However, determining how the network can be optimally designed and then dynamically reconfigured in the event of arbitrarily widespread and catastrophic failure to maximize survivability, particular of critical services, is still very much an open issue, and is the topic of my proposed research. This research will extend current survivability principles to allow for disaster recovery where possible, and develop new methods to deal with aspects of disaster recover that cannot be accommodated with conventional protocols. More to the point, we will seek to develop and demonstrate a suite of survivability mechanisms that are able to recover from arbitrarily widespread network failures and can provide a coordinated network-wide reconfiguration response to maximize survivability of critical services, with minimal additional capital investment. Distributed self-organizing path restoration is particularly well suited to the type of global reconfiguration needed here, and several emerging network survivability approaches (node-inclusive span restoration, failure-independent path-protecting p-cycles, node-encircling p-cycles) also show promise in this regard. The focus of this research will involve the modification of those existing protocols and development of new ones to allow for a coordinated network-wide response to a catastrophic failure, particular so that critical or emergency communication are maximally restorable.