Conventional end-to-end protection switching schemes - e.g., Dedicated Path Protection (DPP) and Shared Path Protection (SPP) - offer established connections a 100% protection against any single failure in the network. However, in the presence of double failures, the performance of these schemes is not fully pre-determined as it depends upon the failure pattern, network topology, and location of both source and destination of the affected connections. In the paper a methodology is presented that provides a rigorous approach to provisioning connections in the presence of any (multi) link failure pattern. The novelty of the proposed methodology is the introduction of the connection's required Maximum Downtime Ratio (or MDR) into the network protection design process. The connection's reliability degree - expressed by means of the MDR - is now formulated in probabilistic terms, and must be met independently of the topology and location of source and destination. With the proposed methodology, the connection's required MDR is considered as an input parameter. The output is the protection scheme of interest that meets such degree of reliability (MDR) with the minimal use of network resources. To best understand how the proposed methodology works, a study is conducted in which a modified SPP scheme is proposed. To cover a significantly wide range of MDR values the SPP scheme is modified to cope with double link failures. The modified SPP is compared against the DPP in terms of amount of required network resources (cost) versus the connection's required MDR. Results show that with the proposed methodology, it is possible to design and select the most cost-effective protection scheme over a large variety of desired connection's MDR values.
展开▼