检索结果-内蒙古大学图书馆

2006 ACM/IEEE Conference on Supercomputing, SC'06

作者： Sur, Sayantan Koop, Matthew J. Panda, Dhabaleswar K. Network-Based Computing Laboratory Department of Computer Science and Engineering Ohio State University

ISBN: (纸本)0769527000

InfiniBand is an emerging HPC interconnect being deployed in very large scale clusters, with even larger InfiniBand-based clusters expected to be deployed in the near future. The Message Passing Interface (MPI) is the programming model of choice for scientific applications running on these largescale clusters. Thus, it is very critical for the MPI implementation used to be based on a scalable and high-performance design. We analyze the performance and scalability aspects of MVAPICH, a popular open-source MPI implementation on InfiniBand, from an application standpoint. We analyze the performance and memory requirements of the MPI library while executing several well-known applications and benchmarks, such as NAS, SuperLU, NAMD, and HPL on a 64-node InfiniBand cluster. Our analysis reveals that latest design of MVAPICH requires an order of magnitude less internal MPI memory (average per process) and yet delivers the best possible performance. Further, we observe that for these benchmarks and applications evaluated, the internal memory requirement of MVAPICH remains nearly constant at around 5-10 MB as the number of processes increase, indicating that the MVAPICH design is highly scalable. © 2006 IEEE.

关键词： Interfaces (computer)

来源：评论

学校读者我要写书评

暂无评论

MDMA read based Rendezvous Protocol for MPI over InfiniBand: Design alternatives and benefits

MDMA read based Rendezvous Protocol for MPI over InfiniBand:...

引用

2006 ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming, PPOPP'06

作者： Sur, Sayantan Jin, Hyun-Wook Chai, Lei Panda, Dhabaleswar K. Network-Based Computing Laboratory Department of Computer Science and Engineering Ohio State University

ISBN: (纸本)1595931899

Message Passing Interface (MPI) is a popular parallel programming model for scientific applications. Most high-performance MPI implementations use Rendezvous Protocol for efficient transfer of large messages. This protocol can be designed using either RDMA Write or RDMA Read. Usually, this protocol is implemented using RDMA Write. The RDMA Write based protocol requires a two-way handshake between the sending and receiving processes. On the other hand, to achieve low latency, MPI implementations often provide a polling based progress engine. The two-way handshake requires the polling progress engine to discover multiple control messages. This in turn places a restriction on MPI applications that they should call into the MPI library to make progress. For compute or I/O intensive applications, it is not possible to do so. Thus, most communication progress is made only after the computation or I/O is over. This hampers the computation to communication overlap severely, which can have a detrimental impact on the overall application performance. In this paper, we propose several mechanisms to exploit RDMA Read and selective interrupt based asynchronous progress to provide better computation/communication overlap on InfiniBand clusters. Our evaluations reveal that it is possible to achieve nearly complete computation/communication overlap using our RDMA Read with Interrupt based Protocol. Additionally, our schemes yield around 50% better communication progress rate when computation is overlapped with communication. Further, our application evaluation with Linpack (HPL) and NAS-SP (Class C) reveals that MPI_Wait time is reduced by around 30% and 28%, respectively, for a 32 node InfiniBand cluster. We observe that the gains obtained in the MPI_Wait time increase as the system size increases. This indicates that our designs have a strong positive impact on scalability of parallel applications. Copyright © 2006 ACM.

关键词： network protocols

来源：评论

学校读者我要写书评

暂无评论

Scaffolding supervision and behavior understanding

Scaffolding supervision and behavior understanding

引用

9th Joint Conference on Information Sciences, JCIS 2006

作者： Chen, Yung-Hui Deng, Lawrence Y. Department of Computer Information and Network Engineering Lunghwa University of Science and Technology Taiwan Department of Computer Science and Information Engineering St. John's University Taiwan

ISBN: (纸本)9078677015

Since the instructional activities are implemented on Cyberspace, how to control behaviors of students and to increase the degree of communication awareness has been a challenging issue. This paper presents an advanced Petri Net model to analyze and supervise of the web-based multiple participants in virtual environment. We proposed a system based on the scaffolding theory. Behaviors of students are supervised by an intelligent control system, which is programmed by the instructor under our generic interface. Problems of providing the multi-user interaction on the Web and the solutions proposed by the Petri Net model are fully elaborated here. This paper can be used as a basic/fundamental research framework and tools to study and understand the characteristics of e-learning and to explore its optimal education application.

关键词： Distance education

来源：评论

学校读者我要写书评

暂无评论

Facilitating metadata creation - The MINE metadata wizard for IEEE LOM

Facilitating metadata creation - The MINE metadata wizard fo...

引用

作者： Lin, H.W. Shih, Timothy K. Multimedia Information Network Lab Department of Computer Science and Information Engineering Tamkang University Taiwan Taiwan Department of Computer Science and Information Engineering Tamkang University Taiwan Taiwan IEEE ACM Department of Computer Science and Information Engineering Tamkang University

One stumbling block to the establishment of distance learning programs is on the time consuming tasks to create high quality digital contents that can be used in instruction. In the literature, reusability of learning objects is addressed as a potential solution toward the sharing of high quality contents. To find suitable contents while one is designing the course material, the IEEE LOM provides a generic and standard mechanism to enable a common representation and vocabulary for the description of learning objects. However, it is unlikely that an individual will fill out the 77 items of metadata defined by the IEEE LOM. Nevertheless, it is possible to compute the relationship among these 77 items, and to provide an automatic metadata generation mechanism to reduce the load of a content creator. This paper proposed such a mechanism, by means of using user profiles, system extracted information, interactive questions, and influence rules. We present the mechanism, as well as a developed system, called the MINE Metadata Wizard, which is integrated with a SCORM-compliant authoring tool that we have developed. Interested readers are welcome to download the tool under http://***. ***/.

关键词： Metadata

来源：评论

学校读者我要写书评

暂无评论

Using description logics for network vulnerability analysis

Using description logics for network vulnerability analysis

引用

International Conference on networking, International Conference on Systems andInternational Conference on Mobile Communications and Learning Technologies, ICN/ICONS/MCL'06

作者： Zakeri, Reza Jalili, Rasool Abolhassani, Hassan Shahriari, Hamid Reza Sharif University of Technology Computer Engineering Department Network Security Center Semantic Web Laboratory

ISBN: (纸本)0769525520

The distributed nature and complexity of computer networks and various services provided via them, makes the networks vulnerable to numerous attacks. The TCP/IP presumptions which are based on using this protocol to provide a simple, open communication infrastructure in an academic and collaborative environment, causes this protocol lack of built-in mechanisms for authentication, integrity and privacy. Even though in the last few years a more systematic approach to TCP/IP network security problem has been followed, a formal approach to this problem is lacking. In this paper, we propose using Description Logics as a formal model which could be used to analyze TCP/IP networks against attacks. Moreover we have presented a case study which models one of famous attack categories, i.e. Man in the Middle (MTM). This model lets automatically verify and proof network safety or vulnerability against this attack. © 2006 IEEE.

关键词： Telecommunication networks

来源：评论

学校读者我要写书评

暂无评论

A hardware implementation of lightweight block cipher for ubiquitous computing security

引用

10th International Conference on Knowledge-Based Intelligent Information and engineering Systems, KES 2006

作者： Jong, Sou Park Kim, Sung-Hwan Dong, Seong Kim Network Security Lab. Computer Engineering Department Hankuk Aviation University Korea Republic of

ISBN: (纸本)3540465359

Ubiquitous Computing applications have resource constraints and conventional block ciphers are infeasible to meet their requirements. This paper presents a hardware design and implementation of lightweight block cipher based on Reversible Cellular Automata (RCA), which is able to generate high pseudo random patterns with low computational overheads. The prototype implementation shows that the total number of gate is less than 3k and operates for 400 clock cycles in 82.156 MHz and it outperforms AES and NTRU. In addition, the cryptanalysis including Strict Avalanche Criterion (SAC), Differential Cryptanalysis (DC) and Linear Cryptanalysis (LC) to our implementation is satisfied. © Springer-Verlag Berlin Heidelberg 2006.

关键词： Security of data

来源：评论

学校读者我要写书评

暂无评论

Dynamic Bandwidth Allocation and Rate Coordination for DiffServ Environment

引用

Wuhan University Journal of Natural Sciences 2006年第1期11卷 248-252页

作者： CHEN Gang WU Guo-xin ZHANG San-feng YOU Feng-qin Department of Computer Science and Engineering Southeast University Nanjing 210096 Jiangsu China Key Laboratory of Computer Network and Information Integration of MOE Southeast University Nanjing 210096Jiangsu China

This paper presents a novel model for dynamic bandwidth allocation and rate coordination based on DiffServ and a bandwidth broker（BB）. In this model, assignment of bandwidth was made according to a periodic trace of network characteristics per application. And adjustment of transfer rate was accomplished through negotiation with applications by a bandwidth agent. This model was evaluated using network simulator 2 （NS-2）, and distinct improvements were found in respects of delay and packet loss of overall network and single flow. Finally, the model was suggested to be leveraged to multimedia applications with properties of lower delay and lower packet loss.

关键词： QoS bandwidth broker（BB） dynamic bandwidth allocation rate coordination

来源：评论

学校读者我要写书评

暂无评论

Cost management based security framework in mobile ad hoc networks

引用

Journal of Zhejiang University-Science A(Applied Physics & engineering) 2006年第4期7卷 493-501页

作者： YANG Rui-jun XIA Qi PAN Qun-hua WANG Wei-nong LI Ming-lu Department of Computer Science ＆ Engineering Shanghai Jiao Tong University. Shanghai 200030. China Network Information Center Shanghai Jiao Tong University Shanghai 200030. China

Security issues are always difficult to deal with in mobile ad hoe networks. People seldom studied the costs of those security schemes respectively and for some security methods designed and adopted beforehand, their effects are often investigated one by one. In fact, when facing certain attacks, different methods would respond individually and result in waste of resources. Making use of the cost management idea, we analyze the costs of security measures in mobile ad hoc networks and introduce a security framework based on security mechanisms cost management. Under the framework, the network system＇s own tasks can be finished in time and the whole network＇s security costs can be decreased. We discuss the process of security costs computation at each mobile node and in certain nodes groups. To show how to use the proposed security framework in certain applications, we give examples of DoS attacks and costs computation of defense methods. The results showed that more secure environment can be achieved based on the security framework in mobile ad hoc networks.

关键词： network attacks Mobile ad hoc Cost management Security framework

来源：评论

学校读者我要写书评

暂无评论

Shared receive queue based scalable MPI design for InfiniBand clusters 20

Shared receive queue based scalable MPI design for InfiniBan...

引用

20th IEEE International Parallel and Distributed Processing Symposium, IPDPS 2006

作者： Sur, Sayantan Chai, Lei Jin, Hyun-Wook Panda, Dhabaleswar K. Network-Based Computing Laboratory Department of Computer Science and Engineering Ohio State University United States

ISBN: (纸本)1424400546

Clusters of several thousand nodes interconnected with InfiniBand, an emerging high-performance interconnect, have already appeared in the Top 500 list. The next-generation InfiniBand clusters are expected to be even larger with tens-of-thousands of nodes. A high-performance scalable MPI design is crucial for MPI applications in order to exploit the massive potential for parallelism in these very large clusters. MVAPICH is a popular implementation of MPI over InfiniBand based on its reliable connection oriented model. The requirement of this model to make communication buffers available for each connection imposes a memory scalability problem. In order to mitigate this issue, the latest InfiniBand standard includes a new feature called Shared Receive Queue (SRQ) which allows sharing of communication buffers across multiple connections. In this paper, we propose a novel MPI design which efficiently utilizes SRQs and provides very good performance. Our analytical model reveals that our proposed designs will take only 1/10ththe memory requirement as compared to the original design on a cluster sized at 16,000 nodes. Performance evaluation of our design on our 8-node cluster shows that our new design was able to provide the same performance as the existing design while requiring much lesser memory. In comparison to tuned existing designs our design showed a 20% and 5% improvement in execution time of NAS Benchmarks (Class A) LU and SP, respectively. The High Performance Linpack was able to execute a much larger problem size using our new design, whereas the existing design ran out of memory. © 2006 IEEE.

关键词： Interconnection networks

来源：评论

学校读者我要写书评

暂无评论

Instant management infrastructure - Through network management support systems and software agents coordination

Instant management infrastructure - Through network manageme...

引用

International Conference on e-Business (ICE-B 2006)

作者： Folha, Jose A. Marques, Bruno F. Oliveira, Jose A. Coelho, Paulo M. Oliveira, Raul F. Responsible Network and Systems Management Superior School of Biotechnology Portuguese Catholic University Portugal Electrical and Computer Engineering's Department Engineering Faculty of University of Porto Portugal Electrical Engineering Department Polytechnic Institute of Viseu's Superior School of Technology Portugal Computer Science Department Polytechnic Institute of Viseu's Superior School of Technology Portugal

ISBN: (纸本)9728865627

Enterprises need to automate manual and routine aspects of IT infrastructure management. In this paper a new concept that integrates different autonomous and management level applications through Instant Messaging protocol (XMPP) is introduced. This is achieved by means of a common management information model implemented in LDAP. A content language implemented in XML is also described, allowing autonomous application integration to carry out management tasks dynamically.

关键词： Internet services networking management applications LDAP basic Internet protocols software agents instant messaging

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：