GMPLS网络中多约束QoS路由的预计算方法

GMPLS(generalized multiprotocol label SWitching)网络中的多约束QoS路由问题是要在诸如带宽、代价和延迟的约束条件下找到一条优化的路径.这个问题通常被认为是一个NP-完全问题.在研究共享风险链路组具有的启发信息的基础上,提出了一种具有共享风险链路启发信息的多约束预计算算法.该算法包含预计算和搜索两个部分.预计算主要是能创建和更新每个节点上的路由表.而后,搜索部分则可以在层次化的结构中选择满足约束条件的优化的路径.大量仿真数据表明,相应的方法能够取得满意的结果,可以有效地解决GMPLS网络中多约束的QoS路由问题.

Precomputation for Multi-Constrained QoS Routing in GMPLS Networks

Multi-Constrained QoS routing in GMPLS (generalized multiprotocol label switching) network is to find an optimal path satisfying several constraints, such as bandwidth, cost and delay. The problem has been considered as a NP-Complete problem. Based on SRLG heuristic information, the paper provides a MPAS algorithm (Multi-constrained Precomputation Algorithm with SRLG), which includes the precomputation and searching procedures. The precomputation i s able to create and update the routing tables in each node. Then, the searching procedure can select an optimal path satisfying several constraints in the hierarchical architecture. The results of extensive simulation based on the self-similar traffic show that the corresponding methods can achieve satisfactory performance and efficiently solve the problem of multi -constrained QoS routing in GMPLS network.