Mean waiting time evaluation of packet switches for centrally controlled PBX''s

The performance analysis of packet switched networks usually does not account for the loss of transmission bandwidth caused by the switch processing time. In wide area networks, where transmission channels are of the order of 64 kbps or less, this is a valid approach since switch processing time is negligible compared to average transmission times for packets. However, in Local Area Networks (LANs), the transmission channels are usually of the order of 1 mbps or greater and thus when evaluating the performance of a LAN that is based on a packet switch, it is necessary to consider the effect of the switch processing time which is now comparable to that of the average packet's transmission time. This paper investigates the effect of this processing time on network performance, and an exhaustive cyclic queueing algorithm to minimise this effect on overall network performance is discussed. An approximate analytical technique for evaluating the performance of such a LAN is derived. This technique makes it possible to do interactive calculations which can be used in the provisioning and configuring of such networks.     

Fat-tree routing and node ordering providing contention free traffic for MPI global collectives

As the size of High Performance Computing clusters grows, so does the probability of interconnect hot spots that degrade the latency and effective bandwidth the network provides. This paper presents a solution to this scalability problem for real life constant bisectional-bandwidth fat-tree topologies. It is shown that maximal bandwidth and cut-through latency can be achieved for MPI global collective traffic. To form such a congestion-free configuration, MPI programs should utilize collective communication, MPI-node-order should be topology aware, and the packet routing should match the MPI communication patterns. First, we show that MPI collectives can be classified into unidirectional and bidirectional shifts. Using this property, we propose a scheme for congestion-free routing of the global collectives in fully and partially populated fat trees running a single job. The no-contention result is then obtained for multiple jobs running on the same fat-tree by applying some job size and placement restrictions. Simulation results of the proposed routing, MPI-node-order and communication patterns show no contention which provides a 40% throughput improvement over previously published results for all-to-all collectives.     

基于网上开放虚拟实验室的研究与设计

首先对网上开放虚拟实验室的基本概念、体系结构和管理模式进行介绍,分析了设计网上开放虚拟实验室的关键技术。最后设计和实现了网上开放虚拟实验室。     

基于Widrow-Hoff神经网络的多指标推荐算法

为解决传统的协同过滤推荐算法不能综合运用多个指标进行推荐的问题,通过引入多指标评分的概念对标准的协同过滤推荐算法进行扩展,提出一种基于Widrow-Hoff神经网络的多指标推荐算法。利用Widrow-Hoff最小二乘法自适应算法在进行系统辨识时的高精度拟合特性,提出一种基于Widrow-Hoff最小二乘法算法的用户偏好特征向量计算方法。利用用户偏好特征向量和空间距离矩阵度量用户相似度,以定位邻居集并为用户推荐最优项目。实验结果表明,本文算法可提高推荐精度,改进推荐质量。     

PGL-CIMS的设计与开发

本文介绍了平果铝业公司CIMS的系统结构、网络结构、软件结构、关键技术等一系列技术问题。     

Win2003 Server中用户的批量建立和管理

使用一种简单有效的批处理文件方法解决大量用户建立的问题,简化了管理任务繁重,重复性工作多,出错几率增大的网络管理工作,极大的提高了个人和单位的工作效率,减少了错误发生,对于在网络出现不可恢复故障后及时重新构建用户方面具有重要的意义。     

水源热泵中央空调控制与网络监控系统

水源热泵中央空调控制与网络监控系统主要是针对现代使用的一种新型高效、节能、环保的空调进行的原理分析、系统设计和控制设计,实现空调系统的智能化控制,同时该系统运用现代高科技、高速度的网络技术对系统进行网络监控,打破了原有的工业远程监控的模式,在工业自动化领域和计算机网络领域均是一种创新和突破。该系统具有环保、节能、可靠性高、自动化程度高等特点,具有很高的经济效益和社会效益,应用前景广阔。     

桥接局域网第2层拓扑结构的自动发现

With the rapid development and the wide use of Bridged Local Area Networks, the importance of below-OSI-Layer-3 topology comes to be realized. Recent research work has demonstrated that active layer-2 topology can be determined using Address Forwarding Information. The large quantities of address forwarding entries in a typical BLAN, however, make their algorithms both time-consuming and bandwidth-consuming. In view of the above and the fact that transparent bridges are used preponderantly in real networks, we base our algorithm on information supported by Spanning Tree Protocol M1B group, which reduces the cost of acquiring and processing information, and improves the accuracy. Our algorithm relies on standard MIB information and can be used in heterogeneous environments     

利用ACL在校园网内实现安全管理

阐述了ACL(Access Control List)基本原理、分类及简单配置方法,并结合实际案例进行了分析并配置,说明如何在校园网中使用ACL进行网络层访问权限控制,提高网络整体性能和安全.     

BP神经网络权值初始值与收敛性问题研究

BP神经网络的收敛性问题是一直受到广泛关注的问题。本文针对BP网络在运算过程中陷入局部最小区，收敛速度慢的问题，从BP算法的原理出发，讨论了权值初始值对网络训练速度的强烈影响（仿真结果证明了这一点），并提出了解决方法。