GMPLS控制面中的邻居发现和拓扑发现机制的改进研究

Improvement of neighbor discovery and topology discovery mechanisms in GMPLS

针对GMPLS控制面中邻居发现和拓扑发现机制的缺陷,尝试从2个方面对其加以改进。一方面,现有LMP(链路管理协议)实现邻居发现机制时,会出现端口不一致的现象,导致不能绑定链路,通过增加消息使之可以顺利进行链路绑定,从而减少了链路维护的开销;另一方面,使用OSPF协议进行拓扑发现时,由于现有的LSA(链路状态公告信息)缺少对LSA更新时间、链路消耗等的支持,造成网络的链路拥塞、带宽利用率不高、节点对故障敏感性差等问题,通过扩展sub-TLV的类型,增加了对LSA更新时间、链路消耗、风险共享和层次属性的规定,从而可以提高链路利用率,减少网络拥塞。

Aiming at the limitation of neighbor and topology discovery mechanisms of GMPLS, this paper attempts to improve them in two ways. On the one hand, when LMP （link management protocol） implements neighbor discovery mechanism, the port disaccord may cause the links cannot be bound. By increasing the messages, the problem is solved and the maintenance cost of links is reduced. On the other hand, current LSA does not support the LSA updating time and link consuming, and this will bring about link congestion and the low bandwidth utilization. In this paper, the four types of sub-TLV are extended to support the updating time of LSA, link consuming, risk sharing and layer attributes for improving the link utilization and decreasing the network congestion.