Integrating network quality, performance and cost control through reliability analysis in optical network design

Abstract

Optical network transports are traditionally designed with ring protection and/or dedicated (1+1) protection as they provide the best reliability for the networks. However, with there is a need to reduce capital expenditure, to improve the efficiency of resources usage for protection, and to keep the performance and the quality of service similar to the ring networks. For this network designers need to have the capability to assess the end-to-end reliability of various design options in order to make appropriate intelligent and economical decision on network design options. One way to achieve this is to implement shared mesh protection, no protection or preemptible services for traffics demands that do not require dedicated protection or 50 ms recovery time achieved solely with dedicated (1+1) or ring protection. However, designers understand that there will be sacrifices in the reliability levels in doing so, but there has not been a systematic methodology of quantifying the reliability of networks with shared protections. This work introduces a new framework for a reliability-based optical network redesign and reoptimization to improve the quality and performance of traffic transport. This framework integrates the end-to-end reliability assessment of the network, which used to be conducted at the end stage of network design, into the network design process. A new hybrid analytical and simulation methodology has been developed to quantify the reliability of various network elements and protection schemes. The reliability profile of the network provides valuable information for network designers or service providers to implement their design more economically, via the use of shared protection, while maintaining an acceptable level of reliability. The acceptable level of reliability could be realized through the redesign and reoptimization of the failing parts of the network; this will improve both quality and performance (BER) while maintaining lower economics than dedicated protection design. This work illustrates three of the many benefits provided by this approach: (1) the use of the reliability profile to compare real optical network designs with different protection schemes; (2) the use of the reliability profile to identify and redesign the vulnerable demand pairs routing; (3) making decision on trade-offs between cost of improving the network reliability and offering multilevel quality of service and performance at the network transport level.