Research On Multi-node Initiation Restoration Mechanism In WDM Networks With Traffic Grooming Capacity

With the development of wavelength division multiplexing (WDM) technology, the resource utilization rate of network bandwidth has improved immensely. To improve the resource utilization rate and reduce the cost of optical network, an efficient approach is to pack low-speed traffic connection requests onto high-speed wavelength channels, which is referred to traffic grooming. As the capacity of WDM grooming network is huge, network failures such as link fiber fractures will cause catastrophic consequences and a large amount of business interruption. Therefore, the research of survivable traffic grooming problem is important.In order to solve the survivable traffic grooming problem, this thesis studies the dynamic restoration mechanisms in WDM networks with traffic grooming capacity. A complete dynamic restoration mechanism is composed of rerouting algorithm and restoration protocol. Firstly, this thesis addresses a dynamic grooming algorithm (DGAL) based on link bound auxiliary graph (LBAG), and uses it as the rerouting algorithm. Secondly, this thesis presents two kinds of restoration protocols, bi-node initiation restoration protocol (BNRP) and multi-node initiation restoration protocol (MNRP), and employs them in WDM grooming networks. The bi-node initiation restoration mechanism has been proposed against single-link failure. According to the different restoration granularities, this thesis designs the connection and lightpath bi-node initiation restoration mechanisms, C-BNRM and L-BNRM respectively. Afterwards, the C-BNRM is extended and the multi-node initiation restoration mechanism against single-link failure (S-MNRM) is proposed. Considering the expending of network scale, there is a possibility of double-link failures in the WDM grooming networks. This thesis finally designs the bi-node initiation restoration mechanism and multi-node initiation restoration mechanism against double-link failures, D-BNRM and D-MNRM respectively.This thesis finally makes the simulation and implementation of these dynamic restoration mechanisms which proposed in this thesis. Simulation results show that C-BNRM provides more flexibility. C-BNRM has a better performance of failure restoration rate and resource utilization rate than L-BNRM, and in terms of failure restoration time, it also achieves a shorter performance. The S-MNRM is the extension of C-BNRM. This thesis makes a comparison among S-MNRM, C-BNRM, source node initiation restoration mechanism (SIRM) and dynamic restoration algorithm without traffic grooming (DPR). The results show that the S-MNRM has a best recovery performance. The S-MNRM increases the number of restoration process initiation node, and shortens the transmission delay of failure message, which reduces the failure restoration time. Besides traffic grooming technology helps to improve the failure restoration rate as well as to reduce failure restoration time. From the comparison between D-BNRM and DMNRM, the D-MNRM reduces the failure restoration time without increasing failure restoration rate, resource utilization rate and blocking rate. It can be concluded that the D-MNRM has a better performance than D-BNRM.