摘要
数据中心网络(DCN)地址配置过程中可能存在配置冲突,利用图同构思想进行冲突检测是一种有效而且可行的方法。从分析冲突类型入手,提出了基于子图同构的冲突检测方法,并对O2算法中基于SPLD产生的关键对的计算进行优化,减少潜在误报;在子图同构的x跳数选择上,探讨给出适当的跳数大小,利用二分法寻找同构子图;对不引起度改变的错误连线故障,通过记录两端节点的方式记录错误连线;改进后的方法一定程度上提升了冲突检测方法的性能,达到了预期效果。
Data center network address configuration may exist configuration conflicts,it is an effective and feasible method to use graph isomorphism ideological for conflict detection. Analysis of the type of conflict first,proposed a method of detection conflict based on subgraph isomorphism,and optimized the O2 algorithm that key generated based on SPLD,reduced the potential for false positives ;Subgraph isomorphism in the number of hops the choice of x, give the appropriate number of hops of size,use dichotomy looking for subgraph isomorphism;On the degree of change does not cause the error connection failed, the way by recording both ends of the nodes record the miswirings;Improved method to some extent enhances the performance of conflict detection method to achieve the desired effect.
出处
《计算机技术与发展》
2011年第11期77-80,共4页
Computer Technology and Development
基金
国家自然科学基金(90604006)
关键词
数据中心网络
冲突检测
子图同构
data center network
conflict detection
subgraph isomorphism
