摘要
IPv6具有巨大的地址空间,未来要面对的将会是海量IPv6路由表,而且128位的IPv6地址比IPv4需要更多的访存数。算法针对IPv6路由查找问题中的这两个难点,提出利用B树高度较低的优良性质,将前缀转化为范围表保存在B树中,并在结点内部利用分段范围比较树算法来减少访存次数和空间耗费。理论分析和实验表明,该算法能够以很好的性能支持IPv6海量路由表的查找。
The IPv6 routing lookup algorithms need to process huge route tables in the future owing to the huge address space of IPv6, and each lookup needs more memory accesses than IPv4 algorithms because of the 128 bits address. To solve these two difficult problems, this algorithm converts the prefix into ranges and stores them in a B-tree, then uses range fragment tree in the nodes to reduce the memory access and the storage requirement. Theoretical analysis and the experimental results indicate that the algorithm can support the high performance lookup for huge IPv6 route tables.
出处
《国防科技大学学报》
EI
CAS
CSCD
北大核心
2005年第5期18-24,共7页
Journal of National University of Defense Technology
基金
国家自然科学基金项目(90104001)
国家重点基础研究发展计划项目(2003CB314802)
国家863高技术研究发展计划基金项目(2003AA115130)