Adopting the borrowed address algorithm can decrease the orphan nodes in ZigBee networks that use distributed address assignment mechanism (DAAM). The existing borrowed address algorithms can increase the success ra...Adopting the borrowed address algorithm can decrease the orphan nodes in ZigBee networks that use distributed address assignment mechanism (DAAM). The existing borrowed address algorithms can increase the success rate of address assignment, but they have defects such as greater cost of overhead and time in founding network caused by breaking topology. To solve such problems, we propose an more efficient distributed borrowed address assignment algorithm based on topology maintenance (A2BTM) that has a topology maintenance function. It borrows address firstly from the offspring nodes in the same branch for the orphan nodes and replies distributed the request of the borrowed address message immediately, to maintain the network topology and decrease the overhead and time spent on the mechanism of borrowed address. Theoretical and simulation analyses manifest that AZBTM algorithm outperforms DAAM and its improved algorithms in terms of the overhead and time spent in founding network, on the premise of keeping a higher success rate of address assignment. Furthermore, A2BTM can lessen the influence from detour phenomenon efficiently.展开更多
为了提高大规模网络中别名解析的效率,在用traceroute测量得到的IP级网络拓扑的基础上,提出别名过滤的概念.首先从理论上研究别名关系具有的性质,由此提出处理traceroute数据的3个属性;然后提出并设计了别名过滤算法AF(alias filtering...为了提高大规模网络中别名解析的效率,在用traceroute测量得到的IP级网络拓扑的基础上,提出别名过滤的概念.首先从理论上研究别名关系具有的性质,由此提出处理traceroute数据的3个属性;然后提出并设计了别名过滤算法AF(alias filtering)和别名验证算法VAR(validation of alias relationship).最后,利用CAIDA(Cooperative Association for Internet Data Analysis)的Skitter项目得到的中国、日本、韩国这3个国家因特网的traceroute数据集对上述算法进行了验证分析.结果表明,别名过滤的概念非常重要并且文中提出的算法效率比较高.展开更多
在分析现有P2P(peer to peer)路由算法的基础上,提出了一种基于二阶矩定位、支持多维资源数据描述的高效资源路由算法——FAN(flabellate addressable network)路由算法.FAN算法将节点映射到统一的多维笛卡尔空间,并以节点相对空间原点...在分析现有P2P(peer to peer)路由算法的基础上,提出了一种基于二阶矩定位、支持多维资源数据描述的高效资源路由算法——FAN(flabellate addressable network)路由算法.FAN算法将节点映射到统一的多维笛卡尔空间,并以节点相对空间原点的二阶矩作为子空间管理和资源搜索的依据.FAN路由算法具有O(log(N/k))的高路由效率,在节点加入和退出FAN网络时,更新路由信息的代价为O(klog(N/k)).实验结果表明,FAN路由算法具有路由效率高、维护代价小的优点,是一种P2P环境中支持多维资源数据描述的高效结构化资源路由算法.而且,目前部分基于CAN(content-addressable network)网络的改进算法也可以在FAN网络中适用,并获得更好的路由效率和更低的维护代价.展开更多
针对IP路由器的FIB(Forwarding Information Base)极限问题和分布式IP地址查找中的通信延迟问题,提出了SD-Torus(Semi-Diagonal Torus)直连网络。按照"临近存储"的原则,将路由表划分后存储在每个节点及其邻居节点上,以减少分...针对IP路由器的FIB(Forwarding Information Base)极限问题和分布式IP地址查找中的通信延迟问题,提出了SD-Torus(Semi-Diagonal Torus)直连网络。按照"临近存储"的原则,将路由表划分后存储在每个节点及其邻居节点上,以减少分布式IP地址查找中的通信延迟,提高整体的查找性能。在分析SD-Torus网络拓扑性质的基础上,提出了一种负载均衡的路由算法。基于SystemC的仿真结果表明,使用该结构可以大大降低分布式IP地址查找的通信延迟,提高系统的扩展性。该研究结果可以应用于高性能的分布式IP地址查找。展开更多
基金supported by Natural Science Foundation Project of CQ CSTC (2012jjA40040)the National Natural Science Foundation of China (60972068)
文摘Adopting the borrowed address algorithm can decrease the orphan nodes in ZigBee networks that use distributed address assignment mechanism (DAAM). The existing borrowed address algorithms can increase the success rate of address assignment, but they have defects such as greater cost of overhead and time in founding network caused by breaking topology. To solve such problems, we propose an more efficient distributed borrowed address assignment algorithm based on topology maintenance (A2BTM) that has a topology maintenance function. It borrows address firstly from the offspring nodes in the same branch for the orphan nodes and replies distributed the request of the borrowed address message immediately, to maintain the network topology and decrease the overhead and time spent on the mechanism of borrowed address. Theoretical and simulation analyses manifest that AZBTM algorithm outperforms DAAM and its improved algorithms in terms of the overhead and time spent in founding network, on the premise of keeping a higher success rate of address assignment. Furthermore, A2BTM can lessen the influence from detour phenomenon efficiently.
文摘为了提高大规模网络中别名解析的效率,在用traceroute测量得到的IP级网络拓扑的基础上,提出别名过滤的概念.首先从理论上研究别名关系具有的性质,由此提出处理traceroute数据的3个属性;然后提出并设计了别名过滤算法AF(alias filtering)和别名验证算法VAR(validation of alias relationship).最后,利用CAIDA(Cooperative Association for Internet Data Analysis)的Skitter项目得到的中国、日本、韩国这3个国家因特网的traceroute数据集对上述算法进行了验证分析.结果表明,别名过滤的概念非常重要并且文中提出的算法效率比较高.
文摘针对IP路由器的FIB(Forwarding Information Base)极限问题和分布式IP地址查找中的通信延迟问题,提出了SD-Torus(Semi-Diagonal Torus)直连网络。按照"临近存储"的原则,将路由表划分后存储在每个节点及其邻居节点上,以减少分布式IP地址查找中的通信延迟,提高整体的查找性能。在分析SD-Torus网络拓扑性质的基础上,提出了一种负载均衡的路由算法。基于SystemC的仿真结果表明,使用该结构可以大大降低分布式IP地址查找的通信延迟,提高系统的扩展性。该研究结果可以应用于高性能的分布式IP地址查找。