Optimal Routing Designs in Self-healing Communications Networks

Self-healing communication networks that allow re-routing of demands through switching processes at designated nodes are studied. It is shown how network utilization, demand throughput and reliability of such networks can be studied simultaneously to achieve an optimal design for all three. This is done through a max–min–max multi-commodity network flow formulation of the routing problem in which it is ensured that maximum network throughput is achieved with minimum loss of demands that are blocked due to single switching node failures. It is shown that a node-path linear programming approximation to the multi-commodity network flow formulation solves the problem for medium and large network sizes in moderate computational times.     

Evaluation of Crossbar Architectures for Deadlock Recovery Routers

The performance of interconnection networks is significantly affected by router speed and routing adaptivity, which can be competing factors. To achieve a high-speed, true-fully-adaptive router design, this paper explores the exploitation of dynamic routing behavior identified as routing locality. When routing locality is exploited, it enables the internal crossbar of a router to be partitioned into smaller and faster units without sacrificing true-fully-adaptive routing capabilities. Extensive evaluation of partitioned crossbar designs which exploit routing locality shows that the increased adaptivity offered by deadlock recovery-based routing algorithms can be implemented in routers without sacrificing router speed. The partitioned crossbar designs reduce average message latency by up to 65% and increase maximum network throughput by up to 51%.     

Design of a crossbar VOQ real-time switch with clock-driven scheduling for a guaranteed delay bound

Most commercial network switches are designed to achieve good average throughput and delay needed for Internet traffic, whereas hard real-time applications demand a bounded delay. Our real-time switch combines clearance-time-optimal switching with clock-based scheduling on a crossbar switching fabric. We use real-time virtual machine tasks to serve both periodic and aperiodic traffic, which simplifies analysis and provides isolation from other system operations. We can then show that any feasible traffic will be switched in two clock periods. This delay bound is enabled by introducing one-shot traffic, which can be constructed at the cost of a fixed delay of one clock period. We carry out simulation to compare our switch with the popular iSLIP crossbar switch scheduler. Our switch has a larger schedulability region, a bounded lower end-to-end switching delay, and a shorter clearance time which is the time required to serve every packet in the system.     

基于输入排队的高速交换调度算法研究

高速交换网络一般采用基于定长信元的交换结构,其性能决定于排队策略和信元调度算法.输入排队策略只有和一个有效的调度算法相结合,才能保证交换结构具有良好的吞吐率和时延等性能.主要阐述了基于VOQ的最大数量匹配算法,最大权重匹配算法,稳定结合算法,神经网络算法等输入排队调度算法,分别从技术特点,性能指标和实现复杂度等多个方面进行比较和分析.分析了分布式和集中式两大类调度算法的工作方式,并根据各类算法的特点提出,神经网络算法可以通过定义其优先级函数实现其余各类算法.     

基于分组交换网络的B-Reno性能研究*

通过仿真实验的方法,研究了B-Reno协议在分组交换网络中的性能。实验使用了不同的网络拓扑和链路参数,全面考察了B-Reno的吞吐率和TCP友好性,并将实验结果与其他TCP协议版本相比较。实验结果显示,B-Reno在分组交换网络中能取得高出Reno和New-Reno10%以上的吞吐率,达到与Sack相当的水平;同时,当与Reno（传统TCP）竞争共享信道时,B-Reno也具有良好的TCP友好性。实验结果表明B-Reno在分组交换网络中也能取得良好的性能。     

An artificial neural network approach for routing in distributed computer networks

This paper presents a new approach based on the Hopfield model of artificial neural networks to solve the routing problem in a context of computer network design. The computer networks considered are packet switching networks, modeled as non-oriented graphs where nodes represent servers, hosts or switches, while bi-directional and symmetric arcs represent full duplex communication links. The proposed method is based on a network representation enabling to match each network configuration with a Hopfield neural network in order to find the best path between any node pair by minimizing an energy function. The results show that the time delay derived from flow assignment carried out by this approach is, in most cases, better than those determined using conventional routing heuristics. Therefore, this neural-network approach is suitable to be integrated into an overall topological design process of moderate-speed and high-speed networks subject to quality of service constraints as well as to changes in configuration and link costs.     

Performance enhancement in buffered delta networks using crossbar switches and multiple links

It is shown that the throughput of buffered delta networks can be increased substantially by modifying the structure and operation of their crossbar switches, by using multiple links to connect the switches in the network, or by combining multiple delta networks in parallel. The variations tried in the structure of the crossbar switches are the size of the switch and the number of buffers in each switch, their capacity, and their position in the switch. Among the alternatives mentioned above, the use of two parallel links for each switch connection results in the highest throughput improvement.     

使用多级交换网络进行高性能路由器设计

拓扑结构和路由算法是影响多级交换网络性能的重要因素．在比较多种多级互连拓扑属性的基础上,提出将3D Torus结构应用于大规模交换网络设计．然后针对3D Torus交换网络中报文路由面临的两个关键问题：多路径负载均衡和报文保序,提出一种基于维序的多路径路由算法DMR(dimension-order—based multi—path routing)．该算法可在保证报文顺序的同时在多条路径上平衡负载,提高交换网络吞吐率．最后通过模拟验证了算法的性能,并与维序路由和随机路由算法进行了比较．模拟结果表明,DMR算法的性能优于维序路由算法,能够达到随机路由算法性能水平,同时具有随机路由算法所不具备的报文保序特性．     

基于ATM交换结构的Hopfield神经网络调度算法

针对ATM交换结构，采用输入缓冲和每条入线在同一个时隙内可传送多于一个信元的策略，利用神经网络具有的实时性、高度并行处理能力和易于电路或光电技术实现等特点，提出了一种Hopfield神经网络调度算法。实验仿真比较表明，该方法不但大大提高了吞吐率，消除了队头阻塞造成的性能恶化，而且降低了信元丢失率和较大程度地降低了平均信元时延，提高了ATM交换结构的性能，实现了信元的优化调度。     

一种基于聚类技术的Hopfield网络求解TSP方法

Hopfiled神经网络方法已被广泛用于求解旅行商问题（TSP），但对于解中规模和大规模的TSP，存在效果不理想甚至难以求解的问题。为了较好地解决这个问题，该文提出一种K-Means聚类算法与Hopfield网络方法相结合求解TSP的新方法，先应用聚类算法对所给城市进行聚类以获得几组规模较小的城市，然后对每一组城市应用Hopfield网络方法进行求解，最后把求解后的每组城市连接起来。计算机仿真结果表明，该方法可以获得最优有效解，并且解的质量明显提高，对求解中大规模的TSP比较有效。     

延迟离散Hopfield型神经网络异步收敛性

离散Ｈｏｐｆｉｅｌｄ型神经网络的一个重要性质是异步运动方式下总能收敛到稳定态。同步运行方式下总能收敛到周期不超过２的极限环,它是该模型可以用于联想记忆设计,组合设计计算的理论基础,文中给出了延迟离散Ｈｏｐｆｉｅｌｄ型网络的收敛性定理,在异步运动方式下,证明了对称连接权阵的收敛性定理,推广了已有的离散Ｈｏｐｆｉｅｌｄ型网络的收敛性结果,给出了能量函数极大值点与延迟离散Ｈｏｐｆｉｅｌｄ型网络的稳定态的     

新型互连网络NIN研究

以通信延迟和网络吞吐率这两个重要参数为出发点,以大规模并行处理系统的新型互连网络（ＮＩＮ）进行了理论分析和模拟测试,并与二维网格、环形网以及反图拓扑互连网络进行了比较,结果表明：与其它三种网络相比,ＮＩＮ不仅保持了反图拓扑互连网络可以连接更多处理机的优点,同时还具有较短的通信延迟和较高的网络吞吐率。     

一种新型的联合单播/组播的两级交换结构*

高速交换结构的设计中一般很难以较低的复杂度实现其对组播业务的支持。提出一种联合单播/组播的两级交换结构TSSIUM,该结构通过对Crossbar结构级联一个组播合路结构实现,因此可以极低的复杂度即可实现对组播业务的支持。理论分析和实测数据均表明,该结构在容许的通信量下可以达到100％的吞吐率。     

一种新型的联合单播/组播的两级交换结构

高速交换结构的设计中一般很难以较低的复杂度实现其对组播业务的支持.提出一种联合单播/组播的两级交换结构TSSIUM,该结构通过对Crossbar结构级联一个组播合路结构实现,因此可以极低的复杂度即可实现对组播业务的支持.理论分析和实测数据均表明,该结构在容许的通信量下可以达到100%的吞吐率.     

支持多速率多播的Ad hoc网络资源分配算法

多播流是无线Ad hoc网络中的一种常见和有效的通信形式,尤其是对多媒体业务而言。提出了一种支持多速率多播传输的Ad hoc网络资源分配算法,它通过引入基于价格的流量分配方案来解决多速率多播传输问题,从而能够自适应地分配网络流量,并且最大化网络流的总效用。仿真结果表明,该算法不仅具有良好的收敛性,而且它的多速率多播传输特性使得不同信道条件的用户能获得最佳的网络性能,从而提高了网络吞吐量。     

Throughput Region of Finite-Buffered Networks

Most of the current communication networks, including the Internet, are packet switched networks. One of the main reasons behind the success of packet switched networks is the possibility of performance gain due to multiplexing of network bandwidth. The multiplexing gain crucially depends on the size of the buffers available at the nodes of the network to store packets at the congested links. However, most of the previous work assumes the availability of infinite buffer-size. In this paper, we study the effect of finite buffer-size on the performance of networks of interacting queues. In particular, we study the throughput of flow-controlled loss-less networks with finite buffers. The main result of this paper is the characterization of a dynamic scheduling policy that achieves the maximal throughput with a minimal finite buffer at the internal nodes of the network under memory-less (e.g., Bernoulli IID) exogenous arrival process. However, this ideal performance policy is rather complex and, hence, difficult to implement. This leads us to the design of a simpler and possibly implementable policy. We obtain a natural trade-off between throughput and buffer-size for such implementable policy. Finally, we apply our results to packet switches with buffered crossbar architecture     

Optimization of stochastic networks using simulated annealing for the storage and recalling of compressed images using SOM

In this paper we are studying the optimization of Stochastic Hopfield neural network and the hybrid SOM–Hopfield neural network for the storage and recalling of fingerprint images. The feature extraction of these images has been performed using FFT, DWT and SOM. The feature vectors are stored in the Hopfield network with Hebbian learning and modified Pseudoinverse learning rules. The study explores the tolerance of Hopfield neural networks for reducing the effect of spurious minima in the recalling process by employing the Simulated annealing process. It is observed from the simulations that the capabilities of the Hopfield network can be sufficiently enhanced by making modifications in the feature extraction of the input data. DWT and SOM together can be used to significantly enhance the recall efficiency. The probability of error in recall in the form of spurious minima is minimized by adopting simulated annealing process in the pattern recalling process.     

“Optimal” Hopfield network for combinatorialoptimization with linear cost function

An "optimal" Hopfield network is presented for combinatorial optimization problems with linear cost function. It is proved that a vertex of the network state hypercube is asymptotically stable if and only if it is an optimal solution to the problem. That is, one can always obtain an optimal solution whenever the network converges to a vertex. In this sense, this network can be called the "optimal" Hopfield network. It is also shown through simulations of assignment problems that this network obtains optimal or nearly optimal solutions more frequently than other familiar Hopfield networks.     

A very large-scale switching system by using nested ring-based architecture

In recent years, network applications and hardware technology have been developed in impressive speed. That is, a large-scale network switching system is needed to satisfy all demand among network service providers and population, such as data, voice, image, video on demand, videoconferencing, telecommunications, remote control and teaching, etc. A general large-scale network switching system cannot fulfill various application needs, such as the correctness of data transmission and the capacity of extension for input/output port of switching system. In this paper, we propose a nested ring-based architecture to build a very large-scale network switching system. In order to satisfy the various network application needs, a nested ring-based architecture is designed as a switching element. It can make input data exchange fast to the destined output port, and input/output port of switching system can easily be extended up to 100 ports or 1000 ports to construct a very large-scale network switching system. The simulation results show that a better performance can be achieved in data transmission, delay, throughput and extensibility for the proposed system.     

基于神经计算的宽带网的动态路由分配

论文对宽带网的动态路由分配问题进行了讨论,提出了一种基于Hopfield神经网络的算法,它具有能高速实时实现的优点及对网络状态和拓扑结构变化的适应性。计算机模拟表明此算法的有效性。