Also we design a link state database storage scheme of distributed storage and redundance backup which ease the implementation of the routing table computation in routing node and synchronization between neighbor routers 同时对域内链路状态数据库采用了分布式存储、冗余备份的方案，易于实现各节点上的路由表计算及路由器间的同步。
The efficiency of the algorithm devised originally for routing tables allows for a linear increase in memory needs : if the number of transition rules rises from one to 10 , memory demands go up by a comparable factor 这种源自路由表的演算法极具效率，对记忆体的需求即使呈线性成长也无所谓：如果规则数量由1增加到10 ，记忆体需求量也会大致成长10倍。
3 . the multiple qos adaptive routing algorithm has been proposed . it can refresh the router table by the previous delay information to find a suitable path with the constraints of the bandwidth and the delay - jitter 该算法直接利用前一周期的时延信息来更新路由表，以作为当前寻找路径的依据，在满足带宽和时延波动约束条件下，使数据传送过程时延最小化。
This paper started with discussing the developing of internet . first , we have discussed several questions which ipv4 is facing , such as the limit of ipv4 address space , the increasing size of the routing tables , the absence of qos 本文从internet的发展谈起，讨论了目前ipv4所面临的几个主要问题： ( 1 )地址空间的不足； ( 2 )路由表过于庞大； ( 3 )缺乏服务质量。
If the routing table contains routes that match the service name and have different service broker identifiers , arbitrarily pick a service broker identifier and match only the routes that use that service broker identifier 如果路由表包含匹配服务名称但service broker标识符不同的多个路由，则任意选取一个service broker标识符，然后只匹配使用该service broker标识符的路由。
In order to avoid too heavy load of the routing node computing routing table in some circumstance , we use a maneuver of load balance between the routing nodes computing routing table and nodes not computing routing table 为防止并行路由表计算中少数情况下可能出现的路由计算节点负荷过重，该策略对各个计算节点及非计算节点进行轮询式查询，若计算节点的负荷过重，则使能非计算节点。
In chapter two we give a multi - dimensional interval routing scheme on hypercube and a routing algorithm based on the compact routing table . we discuss the complexity of the algorithm . in addition , a fault - tolerant algorithm on hypercube is put forward 第二章给出了超立方体结构的一个多维区间路由方法，进而给出了一个基于简明路由表的路由算法，并分析了算法的复杂性，同时还提出了超立方体结构的一个容错路由算法。
Thirdly , we study an efficient parallel routing table computation algorithm , we describe xipeng ' s parallel routing table computation algorithm with formal method , and improve the dividing algorithm of area . the automatic finding and maintenance can be completed in the division 对xipeng提出的并行路由表计算方法进行了形式化的描述，对其发现和维护warr算法进行了改进，实现了路由表计算过程中自动发现及维护warr的功能。
All cards contain one or more lan or wan ports ; a forwarding table , which is a simplified version of the full route table ; the qbrt hardware ; packet - forwarding hardware ; buffering capacity ; and separate interfaces to the route manager and the switch engine on the control board 所有卡都包括：一个或多个局域网或广域网端口、一个转发表（即完整路由表的简化版） 、 qbrt硬件、包转发硬件、缓存容量以与控制板上路由管理器和开关引擎独立的接口。
However , if there is no backup lsp when the primary lsp fails in mpls network , mpls frames cannot be forwarded to destination . therefore , fault recovery has become an important research area in mpls traffic engineering . the main contributions of this paper are as follows : 1 然而,在mpls网络中,当网络发生断线故障时,无法象传统ip网络,数据包可以通过路由表找到别的路径传送到达目的地,在mpls网络中由于断线故障造成一条lsp不通时,如果没有另外一条备份lsp ,将会导致业务流传送断掉。
