检索结果-内蒙古大学图书馆

Reliable Distributed Systems

作者： Chao-Ju Hou K.G. Shin T.K. Tsukada Computer Engineering Division Department of Elect. and Comp. Eng University of Wisconsin Madison WI USA Real-Time Computing Laboratory Department of Elect. Engineering and Comp. Science University of Michigan Ann Arbor MI USA

In recent years a number of load sharing (LS) mechanisms have been proposed or implemented to fully utilize system resources. We design and implement a decentralized LS mechanism based on the Condor package, and give in this paper a description of our design and implementation approaches. Two important features of the design are the use of region-change broadcasts in the information policy to provide each workstation with timely state information at the minimum communication cost, and the use of preferred list in the location policy to avoid task collisions. With these two features, we remove the central manager workstation in Condor, configure its functionalities into each participating workstation, and thus enhance the capability to tolerate single workstation failure and the reliability of Condor. We also discuss the experiments we conduct on the LS mechanism and the observations we obtained from empirical data.< >

关键词： Workstations Packaging Kernel Software packages Chaotic communication Design engineering Broadcasting Costs Availability Delay

来源：评论

学校读者我要写书评

暂无评论

Parallel bidirectional heuristic search on the EM-4 multiprocessor

Parallel bidirectional heuristic search on the EM-4 multipro...

引用

International Symposium on Parallel and Distributed Processing (IPDPS)

作者： A. Sohn M. Sato S. Sakai Y. Kodama Y. Yamaguchi Computer and Information Science Department New Jersey Institute of Technology Newark NJ USA Computer Architecture Section Electro Technical Laboratory Tsukuba Ibaraki Japan Real World Computing Tsukuba Research Center Tsukuba Ibaraki Japan

Solving search problems takes a large amount of computational resources both in terms of execution time and memory usage. This report presents experimental results of Parallel Bidirectional Heuristic Search (PBiHS) on the 80-processor EM-4 multithreaded data-flow multiprocessor. The PBiHS searches from two directions in parallel while search in each direction is also performed in parallel. Important data structures are distributed to all processors to help reduce the execution time of realistic problem sizes down to a few seconds or less. We implement two search problems, the Eight Puzzle and the Tower of Hanoi, and execute on the target multiprocessor. Execution results demonstrate that the Parallel Bidirectional Heuristic Search can solve the tree depth 20-40 of the Eight-Puzzle and the 3-9 disks of the Tower of Hanoi in an optimal or near optimal number of iterations in less than two seconds, is highly scalable as it gives over 40-fold speedup on 80 processors, and yields on the average 10-fold improvement over unidirectional search for the 8-Puzzle while generating a far less number of nodes.< >

关键词： Search problems Poles and towers Artificial intelligence Concurrent computing Data structures Parallel processing information science computer architecture Laboratories Buildings

来源：评论

学校读者我要写书评

暂无评论

ANALYTIC MODELS OF ADAPTIVE LOAD SHARING SCHEMES IN DISTRIBUTED real-time SYSTEMS

引用

IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS 1993年第7期4卷 740-761页

作者： SHIN, KG HOU, CJ Real-Time Computing Laboratory Department of Electrical Engineering and Computer Science University of Michigan Ann Arbor MI USA

In a distributed real-time system, nonuniform task arrivals may temporarily overload some nodes while leaving some other nodes idle. As a result, some of the tasks on an overloaded node may miss their deadlines even if the overall system has the capacity to meet the deadlines of all tasks. In a companion paper [1], we proposed, without any modeling analysis, a decentralized, dynamic load sharing (LS) scheme as a solution to this problem. In this paper, we develop analytic queueing models to comparatively evaluate the proposed LS scheme as well as three other schemes: no LS, LS with random selection of a receiver node, and LS with perfect information. The evolution of a node's load state is modeled as a continuous-time semi-Markov process, where cumulative execution time (CET), rather than the commonly-used queue length (QL), is employed to describe the workload of a node. Not only fundamental differences among the different LS schemes are addressed in the analytic models, but also implementation overheads are taken into account Several metrics relevant to real-time performance are derived from these models: in particular, we evaluate the probability of a task missing its deadline, called the probability of dynamic failure. The proposed scheme is compared against other LS schemes using these performance metrics. The validity of analytic models is checked with simulations. Both analytic and simulation results indicate that by using judicious exchange/use of state information and Bayesian decision mechanism, the proposed scheme makes a significant improvement over other existing LS schemes in minimizing the probability of dynamic failure.

关键词： BAYESIAN ANALYSIS CONTINUOUS-time MARKOV CHAINS DEADLINES DISTRIBUTED real-time SYSTEMS LOAD SHARING LOCATION AND TRANSFER POLICIES PERFORMANCE EVALUATION RANDOM PROBING AND SELECTION

来源：评论

学校读者我要写书评

暂无评论

A SIMPLE DISTRIBUTED LOOP-FREE ROUTING STRATEGY FOR computer-COMMUNICATION NETWORKS

引用

IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS 1993年第12期4卷 1308-1319页

作者： SHIN, KG CHOU, CC Real-Time Computing Laboratory Computer Science and Engineering Division Department of Electrical Engineering and Computer Science University of Michigan Ann Arbor MI USA

The loops resulting from either component failures or Load changes in a computer communication network degrade the performance and the adaptability of conventional distributed adaptive routing strategies, such as ARPANET's previous routing strategy (APRS). In this paper, we develop a distributed loop-free routing strategy by adding only one additional piece of information-the total number of minimum-delay paths-to the commonly used routing messages and tables. Most conventional approaches to the looping problem suffer high overheads in time and space because each message must either include the first several nodes of its path or trace the entire path to detect a loop. By contrast, the proposed routing strategy requires only easily obtainable information, yet removes loops completely. It is far more efficient in both time and space than its conventional counterparts, especiallyTable Ifor sparse computer networks. The correctness of the proposed strategy is proved, and several illustrative e?xamples are given. The performance of this strategy is shown to be always better than, or at least as good as, that of APRS and any multiorder routing strategies, where the order of a routing strategy is determined by the amount of routing information carried in each routing message.

关键词： ARPANET computer COMMUNICATION NETWORKS DISTRIBUTED ADAPTIVE ROUTING LOOP-FREE OR FAULT-TOLERANT ROUTING NETWORK DELAY TABLES STRATEGY ROUTING MESSAGES

来源：评论

学校读者我要写书评

暂无评论

MIXED time-CONSTRAINED AND NON-time-CONSTRAINED COMMUNICATIONS IN LOCAL-AREA NETWORKS

引用

IEEE TRANSACTIONS ON COMMUNICATIONS 1993年第11期41卷 1668-1676页

作者： SHIN, KG ZHENG, Q Department of Electrical Engineering and Computer Science Real-Time Computing Laboratory University of Michigan Ann Arbor MI USA Mitsubishi Electric Research Laboratories Inc. Cambridge MA USA

It is well known that some time-token medium access protocols for local area networks (LANs) like the IEEE 802.4 token bus and the FDDI token ring can guarantee the medium access delay for time-constrained packets. However, a problem which has been largely overlooked is how these protocols can be made to provide a maximum throughput for nontime-constrained packets while guaranteeing the delay bound of time-constrained packets. We first show how the parameters of the IEEE 802.4 token bus and the FDDI token ring can be set to solve the above problem. Then, we design a new timer mechanism for the timed-token protocols which provides the highest guaranteed throughput of nontime-constrained packets among a set of medium access protocols called the token passing protocol, to which most of the existing non-contention LAN protocols belong. We present numerical examples to compare different protocols, all of which have shown the superiority of the proposed protocol to the others.

关键词： MIXING time Packet token bus Local Area Networks Token ring fiber distributed data interface (FDDI) Media Access Protocol

来源：评论

学校读者我要写书评

暂无评论

Assigning ADT modules with random neural networks 26

Assigning ADT modules with random neural networks

引用

26th Hawaii International Conference on System sciences, HICSS 1993

作者： Welch, L.R. Stoyenko, A.D. Chen, S. Real-Time Computing Laboratory Departmental Computer and Information Science New Jersey Institute of Technology NewarkNJ07102 United States

ISBN: (纸本)0818632305

This paper applies random neural networks to the problem of assigning abstract data type modules (ADTs) to the processing elements of parallel computers. Although assignment of tasks has been discussed extensively in the literature, the automatic assignment of ADTs is a relatively new problem, and therefore we describe the problem in detail. This is followed by an introduction to the random neural network model and a presentation of our neural network solution to the assignment problem. We present experimental results comparing our solution to those obtained with random assignment and with a greedy heuristic. The random neural network is found to give significantly better results than the other two approaches in virtually every case. © 1993 IEEE.

关键词： Combinatorial optimization

来源：评论

学校读者我要写书评

暂无评论

Incorporation of optimal timeouts into distributed real-time load sharing 26

Incorporation of optimal timeouts into distributed real-time...

引用

26th Hawaii International Conference on System sciences, HICSS 1993

作者： Hou, C.-J. Shin, K.G. Real-Time Computing Laboratory Department of Electrical Engineering and Computer Science University of Michigan Ann ArborMI48109-2122 United States

ISBN: (纸本)0818632305

This paper addresses the problem of designing and incorporating a timeout mechanism into load sharing (LS) with state-region change broadcasts in the presence of node failures in a distributed real-time system. Failure of a node is diagnosed by the other nodes through communication timeouts, and the timeout period used to diagnose whether a node is faulty or not depends on the dynamic changes in system load, the task attributes at the node, and the state the node was initially in. We formulate the problem of determining the 'best' timeout period Tiout for node i as a hypothesis testing problem, and maximize the probability of detecting node failures subject to a prespecified probability of falsely diagnosing a healthy node as faulty. The parameters needed for the calculation of Tiout are estimated on-line by node i using the Bayesian technique and are piggybacked in its region-change broadcasts. The broadcast information is then used to determine Tiout. If node n has not heard from node t for T Tiout since its receipt of node i's broadcast, it will consider node i failed. On the other hand, each node n also determines its own timeout period Tnout, and broadcasts its state not only at the time of state-region changes but also when it has remained within a broadcast-threshold interval throughout Tnout. Our simulation results show that the LS algorithm which combines the on-line parameter estimation, the timeout mechanism, and a few extra, timely broadcasts can significantly reduce the probability of missing task deadlines. © 1993 IEEE.

关键词： Probability

来源：评论

学校读者我要写书评

暂无评论

Evaluation of load sharing in HARTS while considering message routing and broadcasting 93

Evaluation of load sharing in HARTS while considering messag...

引用

1993 ACM SIGMETRICS Conference on Measurement and Modeling of computer Systems, SIGMETRICS 1993

作者： Shin, Kang G. Hou, Chao-Ju Real-Time Computing Laboratory Department of Electrical Engineering and Computer Science University of Michigan Ann ArborMI48109-2122 United States

In this paper, we apply the load sharing (LS) mechanism proposed in [1, 2] to HARTS, an experimental distributed realtime system [3] currently being built at the real-time computing laboratory of the University of Mic... 详细信息

ISBN: (纸本)0897915801

关键词： Distributed computer systems

来源：评论

学校读者我要写书评

暂无评论

Determination of an optimal retry time in multiple-module computing systems 2

Determination of an optimal retry time in multiple-module co...

引用

2nd International Symposium on Uncertainty Modeling and Analysis, ISUMA 1993

作者： Hou, C.-J.J. Shin, K.G. Real-Time Computing Laboratory Department of Electrical Engineering and Computer Science University of Michigan Ann ArborMI48109-2122 United States

ISBN: (纸本)0818638508

The optimal amount of time used for retrying an instruction on detection of an error in a computing system is usually determined under the assumption that the system is composed of a single module, within which all fault activities are confined until some module-replacement action is taken. The authors consider fault activities in multiple-module systems. They first relax the single-module assumption and model the fault activities in a multiple-module system as a Markov process. The randomization approach is applied to decompose the Markov process into a discrete-time Markov chain subordinated to a Poisson process. Using this decomposition, several interesting measures can be derived such as the conditional probability of successful retry given a retry period and the fact that a non-permanent fault has occurred, and the mean time until which all modules in the system enter a fault-free state. All the measures derived are used to determine, along with the parameters characterizing fault activities and costs of recovery techniques, whether or not retry should be used as a first-step recovery means on detection of an error;and the best retry period subject to a specific probability of successful retry. © 1993 IEEE.

关键词： Markov processes

来源：评论

学校读者我要写书评

暂无评论

Adaptive Deadlock-Free Routing in Multicomputers Using only One Extra Virtual Channel

Adaptive Deadlock-Free Routing in Multicomputers Using only ...

引用

1993 International Conference on Parallel Processing, ICPP 1993

作者： Su, Chien-Chun Shin, Kang G. Real-Time Computing Laboratory Department of Electrical Engineering and Computer Science University of Michigan Ann ArborMI48109-2122 United States

ISBN: (纸本)0849389836

We present three protocols defin ing the relationship between messages and the chan nel resources requested: request-then-hold, requestthen wait, and request-then-relinquish. Based on the three protocols, we develop an adaptive deadlockfree routing algorithm called the SP routing. The SP routing uses shortest paths and is fully-adaptive, so messages can be routed via any of the shortest paths from the source to the destination. Since it is a minimal or shortest routing, the SP routing guar antees the freedom of livelocks. The SP routing is not limited to a specific network topology. The main requirement for an applicable network topology is that there exists a deterministic, minimal, deadlock-free routing algorithm. Most ex isting network topologies are equipped with such an algorithm. In this paper, we present an adaptive deadlock-free routing agorithm for n-dimensional meshes by using the SP routing. The hardware re quired by the SP routing uses only one extra virtual channel as compared to the deterministic routing. © 1993 IEEE.

关键词： Network topology

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：