This In the past decade there has been an increasing need for designs to address the time and cost efficiency issues from various computer network applications such as general IP address lookup and specific network in...This In the past decade there has been an increasing need for designs to address the time and cost efficiency issues from various computer network applications such as general IP address lookup and specific network intrusion detection. Hashing techniques have been widely adopted for this purpose, among which XOR-operation-based hashing is one of most popular techniques due to its relatively small hash process delay. In most current commonly used XOR-hashing algorithms, each of the hash key bits is usually explicitly XORed only at most once in the hash process, which may limit the amount of potential randomness that can be introduced by the hashing process. In [1] a series of bit duplication techniques are proposed by systematically duplicating one row of key bits. This paper further looks into various ways in duplicating and reusing key bits to maximize randomness needed in the hashing process so as to enhance the overall performance further. Our simulation results show that, even with a slight increase in hardware requirement, a very significant reduction in the amount of hash collision can be obtained by the proposed technique.展开更多
A Scalable Multi-Hash( SMH) name lookup method is proposed,which is based on hierarchical name decomposition to aggregate names sharing common prefixes and multiple scalable hash tables to minimize collisions among pr...A Scalable Multi-Hash( SMH) name lookup method is proposed,which is based on hierarchical name decomposition to aggregate names sharing common prefixes and multiple scalable hash tables to minimize collisions among prefixes. We take the component instead of the entire name as a key in the hash functions. The SMH method achieves lookup speeds of 21. 45 and 20. 87 Mbps on prefix table with 2 million and 3. 6 million names,respectively. The proposed method is the fastest of the four methods considered and requires 61.63 and 89.17 Mb of memory on the prefix tables with 2 million and 3. 6 million names,respectively. The required memory is slightly larger than the best method. The scalability of SMH outperforms that of the other two methods.展开更多
In order to classify packet, we propose a novel IP classification based the non-collision hash and jumping table trie-tree (NHJTTT) algorithm, which is based on noncollision hash Trie-tree and Lakshman and Stiliadis p...In order to classify packet, we propose a novel IP classification based the non-collision hash and jumping table trie-tree (NHJTTT) algorithm, which is based on noncollision hash Trie-tree and Lakshman and Stiliadis proposing a 2-dimensional classification algorithm (LS algorithm). The core of algorithm consists of two parts: structure the non-collision hash function, which is constructed mainly based on destination/source port and protocol type field so that the hash function can avoid space explosion problem; introduce jumping table Trie-tree based LS algorithm in order to reduce time complexity. The test results show that the classification rate of NHJTTT algorithm is up to 1 million packets per second and the maximum memory consumed is 9 MB for 10 000 rules. Key words IP classification - lookup algorithm - trie-tree - non-collision hash - jumping table CLC number TN 393.06 Foundation item: Supported by the Chongqing of Posts and Telecommunications Younger Teacher Fundation (A2003-03).Biography: SHANG Feng-jun (1972-), male, Ph.D. candidate, lecture, research direction: the smart instrument and network.展开更多
可变长地址是未来网络领域的重要研究内容之一。针对传统路由查找算法在面向可变长地址时查找效率低的问题,提出一种基于平衡二叉树AVL(Adelson-Velskii and Landis)树和Bloom过滤器的适用于可变长地址的高效路由查找算法,简称为AVL-Bl...可变长地址是未来网络领域的重要研究内容之一。针对传统路由查找算法在面向可变长地址时查找效率低的问题,提出一种基于平衡二叉树AVL(Adelson-Velskii and Landis)树和Bloom过滤器的适用于可变长地址的高效路由查找算法,简称为AVL-Bloom算法。首先,针对可变长地址灵活可变且无界的特点,利用多个片外哈希表分别存储前缀比特位数相同的路由条目及其下一跳信息,同时应用片上Bloom过滤器加速搜索可能匹配的路由前缀;其次,为了解决基于哈希技术的路由查找算法在查找最长前缀路由时需多次哈希对比的问题,引入AVL树技术,即通过AVL树组织每组路由前缀集合的Bloom过滤器及其哈希表,优化路由前缀长度的查询顺序,并减少哈希计算次数进而降低查询时间;最后,在3种不同的可变长地址数据集上将所提算法与METrie(Multi-Entrance-Trie)和COBF(Controlled prefix and One-hashing Bloom Filter)这两种传统路由查找算法进行对比实验。实验结果表明,AVL-Bloom算法的查询时间明显少于METrie和COBF算法,分别减少了将近83%和64%;同时,AVL-Bloom算法在路由表项数变化较大的情况下也能维持稳定的查找性能,适用于可变长地址的路由查找转发。展开更多
使用分布式哈希表(distributed hash table,简称DHT)的应用系统必须在O(1)发现算法和O(logN)发现算法系列中选择适应的DHT协议.但是,不同网络波动程度的应用场景要求理想的DHT协议根据网络波动率能够自适应地调整.提出一种发现算法ROAD(...使用分布式哈希表(distributed hash table,简称DHT)的应用系统必须在O(1)发现算法和O(logN)发现算法系列中选择适应的DHT协议.但是,不同网络波动程度的应用场景要求理想的DHT协议根据网络波动率能够自适应地调整.提出一种发现算法ROAD(routing on active and demand),在延时和波动率之间自适应地调整以提供更好的性能.设计ROAD的关键挑战是构建路由表和降低延时的算法.通过构建加速路由表,加快发现服务的速度,降低消息转发的延时,并通过幂次序组播算法改善对超级点的依赖性.模拟实验显示,与现有DHT算法相比,ROAD维护了一种高效发现延时与波动率的折衷.选择不同质量类型的超级点,ROAD可以扩展成满足不同服务需要的发现机制.展开更多
文摘This In the past decade there has been an increasing need for designs to address the time and cost efficiency issues from various computer network applications such as general IP address lookup and specific network intrusion detection. Hashing techniques have been widely adopted for this purpose, among which XOR-operation-based hashing is one of most popular techniques due to its relatively small hash process delay. In most current commonly used XOR-hashing algorithms, each of the hash key bits is usually explicitly XORed only at most once in the hash process, which may limit the amount of potential randomness that can be introduced by the hashing process. In [1] a series of bit duplication techniques are proposed by systematically duplicating one row of key bits. This paper further looks into various ways in duplicating and reusing key bits to maximize randomness needed in the hashing process so as to enhance the overall performance further. Our simulation results show that, even with a slight increase in hardware requirement, a very significant reduction in the amount of hash collision can be obtained by the proposed technique.
基金sponsored by the National Basic Research Program of China(973 Program)(Grant No.2011CB302605)the National High Technology Research and Development Program of China(863 Program)(Grants No.2011AA010705+5 种基金2012AA0125022012AA012506)the National Key Technology R&D Program of China(Grant No.2012BAH37B01)the National Science Foundation of China(Grant No.6120245761402149)the CNNIC(Grant No.K201211043)
文摘A Scalable Multi-Hash( SMH) name lookup method is proposed,which is based on hierarchical name decomposition to aggregate names sharing common prefixes and multiple scalable hash tables to minimize collisions among prefixes. We take the component instead of the entire name as a key in the hash functions. The SMH method achieves lookup speeds of 21. 45 and 20. 87 Mbps on prefix table with 2 million and 3. 6 million names,respectively. The proposed method is the fastest of the four methods considered and requires 61.63 and 89.17 Mb of memory on the prefix tables with 2 million and 3. 6 million names,respectively. The required memory is slightly larger than the best method. The scalability of SMH outperforms that of the other two methods.
文摘In order to classify packet, we propose a novel IP classification based the non-collision hash and jumping table trie-tree (NHJTTT) algorithm, which is based on noncollision hash Trie-tree and Lakshman and Stiliadis proposing a 2-dimensional classification algorithm (LS algorithm). The core of algorithm consists of two parts: structure the non-collision hash function, which is constructed mainly based on destination/source port and protocol type field so that the hash function can avoid space explosion problem; introduce jumping table Trie-tree based LS algorithm in order to reduce time complexity. The test results show that the classification rate of NHJTTT algorithm is up to 1 million packets per second and the maximum memory consumed is 9 MB for 10 000 rules. Key words IP classification - lookup algorithm - trie-tree - non-collision hash - jumping table CLC number TN 393.06 Foundation item: Supported by the Chongqing of Posts and Telecommunications Younger Teacher Fundation (A2003-03).Biography: SHANG Feng-jun (1972-), male, Ph.D. candidate, lecture, research direction: the smart instrument and network.
文摘可变长地址是未来网络领域的重要研究内容之一。针对传统路由查找算法在面向可变长地址时查找效率低的问题,提出一种基于平衡二叉树AVL(Adelson-Velskii and Landis)树和Bloom过滤器的适用于可变长地址的高效路由查找算法,简称为AVL-Bloom算法。首先,针对可变长地址灵活可变且无界的特点,利用多个片外哈希表分别存储前缀比特位数相同的路由条目及其下一跳信息,同时应用片上Bloom过滤器加速搜索可能匹配的路由前缀;其次,为了解决基于哈希技术的路由查找算法在查找最长前缀路由时需多次哈希对比的问题,引入AVL树技术,即通过AVL树组织每组路由前缀集合的Bloom过滤器及其哈希表,优化路由前缀长度的查询顺序,并减少哈希计算次数进而降低查询时间;最后,在3种不同的可变长地址数据集上将所提算法与METrie(Multi-Entrance-Trie)和COBF(Controlled prefix and One-hashing Bloom Filter)这两种传统路由查找算法进行对比实验。实验结果表明,AVL-Bloom算法的查询时间明显少于METrie和COBF算法,分别减少了将近83%和64%;同时,AVL-Bloom算法在路由表项数变化较大的情况下也能维持稳定的查找性能,适用于可变长地址的路由查找转发。
基金Supported by the National Natural Science Foundation of China under Grant Nos.60433040 60603070 (国家自然科学基金)the China Postdoctoral Science Foundation under Grant No.2005038064 (中国博士后基金)
文摘使用分布式哈希表(distributed hash table,简称DHT)的应用系统必须在O(1)发现算法和O(logN)发现算法系列中选择适应的DHT协议.但是,不同网络波动程度的应用场景要求理想的DHT协议根据网络波动率能够自适应地调整.提出一种发现算法ROAD(routing on active and demand),在延时和波动率之间自适应地调整以提供更好的性能.设计ROAD的关键挑战是构建路由表和降低延时的算法.通过构建加速路由表,加快发现服务的速度,降低消息转发的延时,并通过幂次序组播算法改善对超级点的依赖性.模拟实验显示,与现有DHT算法相比,ROAD维护了一种高效发现延时与波动率的折衷.选择不同质量类型的超级点,ROAD可以扩展成满足不同服务需要的发现机制.
