检索结果-内蒙古大学图书馆

International Conference on Computational Intelligence and Security

作者： Zhensong Liao Hai Jin Deqing Zou Chisong Li Services Computing Technology and System Laboratory Cluster and Grid Computing Laboratory Huazhong University of Science and Technology Wuhan China

Automated trust negotiation (ATN) is an important means to establish trust between strangers through the exchange of digital credentials and mobile access control policies specifying what combinations of credentials a stranger must submit. While ATN provides a good mechanism to share resources in the open and distributed environments, there still exist some pitfalls. A notable problem is that the negotiation process is easy to be interrupted, which leads the negotiation to failure. Thereto, current protection techniques are somehow complex and heavyweight, which greatly limits ATN's applications. To solve these problems, an Email Based ATN Model (EBAM) is proposed. A credential template and an access control policy template are specified to carry information. All the transmitted messages are encrypted by private keys. The data exchange is realized through the negotiators' emails. A use case is given to show how it works, which proves the model sound and reasonable.

关键词： Cryptography Costs Security Access control Protocols Public key Computational intelligence Computers grid computing Mobile computing

来源：评论

学校读者我要写书评

暂无评论

SDPMN: Privacy preserving MapReduce network using SDN

arXiv

引用

arXiv 2018年

作者： Li, He Jin, Hai Services Computing Technology System Lab Cluster Grid Computing Lab School of Computer Science and Technology Huazhong University of Science and Technology Wuhan430074 China

MapReduce is a popular programming model and an associated implementation for parallel processing big data in the distributed environment. Since large scaled MapReduce data centers usually provide services to many users, it is an essential problem to preserve the privacy between different applications in the same network. In this paper, we propose SDPMN, a framework that using software defined network (SDN) to distinguish the network between each application, which is a manageable and scalable method. We design this framework based on the existing SDN structure and Hadoop networks. Since the rule space of each SDN device is limited, we also propose the rule placement optimization for this framework to maximize the hardware supported isolated application networks. We state this problem in a general MapReduce network and design a heuristic algorithm to find the solution. From the simulation based evaluation, with our algorithm, the given network can support more privacy preserving application networks with SDN switches. Copyright © 2018, The Authors. All rights reserved.

关键词： MapReduce

来源：评论

学校读者我要写书评

暂无评论

Vdismodel: A Display Latency Performance Model in Virtualization Environment

Vdismodel: A Display Latency Performance Model in Virtualiza...

引用

Sixth International Conference on Semantics,Knowledge and grids(第六届语义、知识与网格国际会议 SKG 2010)

作者： Bin Zhou Hai Jin Pingpeng Yuan Services Computing Technology and System Lab Cluster and Grid Computing Lab School of Computer Science and Technology Huazhong University of Science and Technology Wuhan 430074 China

With the rapid development of virtualization technology as well as the good performance of computer hardware, virtual machine is beginning to gradually increase the performance of display to meet the needs of desktop users. However, there is a difference between the display mechanism in the virtual environment and the traditional environment. At present, more is focused on how to effectively and pertinently measure and assess the performance of display in virtualization environment. Taking consideration of the characteristics of display in the virtual environment, the virtual machine display performance evaluation system has following advantages: it can select a representative workload running on the VM in Domain U; make use of Tracing and Instrumentation to insert the record in the kernel of VM; log the relevant information; and then finally draw the performance-related parameters by the analysis of the recorder. This paper mainly depicts the underlying implementation details of the virtual display system in Xen and analyzes the display principle of Xen-based virtual machine. It also proposes a display latency performance model in Xen environment, which is named Vdismodel by verifying the reasonableness of the Vdismodel through a series of experiments and comparing the theoretical data and the measured data which are accessed by the virtual machine system performance evaluation.

关键词： Benchmark testing Servers Virtual machining Computational modeling Driver circuits Virtual environment Performance evaluation

来源：评论

学校读者我要写书评

暂无评论

A Map-reduce Method for Training Autoencoders on Xeon Phi

A Map-reduce Method for Training Autoencoders on Xeon Phi

引用

IEEE International Conference on Ubiquitous computing and Communications

作者： Qiongjie Yao Xiaofei Liao Hai Jin Services Computing Technology and System Lab Cluster and Grid Computing Lab School of Computer Science and Technology Huazhong University of Science and Technology Wuhan 430074 China

ISBN: (纸本)9781509001552

The stacked autoencoder is a deep learning model that consists of multiple autoencoders. This model has been widely applied in numerous machine learning applications. A significant amount of effort has been made to increase the size of the deep learning model with respect to the size of the training dataset and the parameter of the model to improve performance. However, training a large deep learning model is highly time consuming. Recent studies have applied the CPU cluster with thousands of machines as well as the single GPU or the GPU cluster, to train large scale deep learning models. As a high-performance coprocessor like GPU, the Xeon Phi can be an alternative tool for training large scale deep learning models on a single machine. The Xeon Phi can be recognized as a small cluster which features about 60 cores, and each core supports four hardware threads. Massive parallelism offsets the low computing capacity of every core, but challenges an efficient parallel autoencoders design. In this paper, we analyze the training algorithm of autoencoders based on the matrix operation and point out the thread oversubscription problem, which results in performance degradation. Based on the observation, we propose our map-reduce implementation of autoencoders on the Xeon Phi coprocessor. Our basic idea is to parallelize multiple autoencoder model replicas with bulk synchronous parallel (BSP) communication model where the parameters are updated after the computations of all replicas are completed. Each thread is responsible for one model replica, and all replicas work together on the same mini-batch. This data parallelism method is suitable for training autoencoders on the Xeon Phi, and can extend to asynchronous parallel training method without thread oversubscription. In our experiment the speedup is four times higher than that of sequential implementation. Enlarging the size of the autoencoder model, our method still gets stable speedup.

关键词： Deep Learning Autoencoder Xeon Phi Mapreduce BSP

来源：评论

学校读者我要写书评

暂无评论

Collaboratively Scheduling to Decrease Inter-AS Traffic in P2P Live Streaming

Collaboratively Scheduling to Decrease Inter-AS Traffic in P...

引用

International Conference on Advanced Information Networking and Applications Workshops (AINAW)

作者： Xuping Tu Hai Jin Xiaofei Liao Weidong Wang Sirui Yang Qi Huang Services Computing Technology and System Laboratory Cluster and Grid Computing Laboratory Huazhong University of Science and Technology Wuhan China

In the existing data fetching algorithms for P2P streaming systems, the peers usually fetch data independently, leading to huge inter-AS traffic (AS: autonomous system). This paper proposes CoFetch, a collaborative fetching algorithm for P2P live streaming systems to decrease inter-AS traffic. In CoFetch, to avoid that all peers fetch the same data from inter-AS connection, instead of aggressively fetching data depending on the peers' own desires, the peers divide requests into multiple steps. In each step, peers mainly and locally share data fetched in previous steps by other collaborating peers and request subset data from inter-AS connections. The analytical results show that, inter-AS traffic in CoFetch is independent of the number of peers in AS but the number of collaborating peers in AS. Furthermore, in each overlay, when we limit the number of collaborating peers in one AS from 5 to 20, the inter-AS traffic is about 1.7-3 times of the streaming rate, regardless of the number of peers.

关键词： Telecommunication traffic Peer to peer computing Network topology Application software Collaborative work International collaboration Processor scheduling grid computing Computer science Electronic mail

来源：评论

学校读者我要写书评

暂无评论

A Flexible two-level mechanism in querying and presenting large-scale historical monitoring data

A Flexible two-level mechanism in querying and presenting la...

引用

Asia-Pacific Conference on Communications

作者： Sheng Di Hai Jin Shengli Li Services Computing Technology and System Laboratory Cluster and Grid Computing Laboratory Huazhong University of Science and Technology Wuhan China

ISBN: (纸本)9781424413737;1424413737

With the increasing scale of data transmitted over Internet, how to effectively query and present mass distributed monitoring data has become a serious problem in distributed monitoring field. Almost all users want to retrieve data as detail as possible; on the other hand, the practical bandwidth and latency limit the data scale in transmission. In order to solve this conflict, we design a flexible two-level mechanism. In our platform with this mechanism, users can customize various retrieving spaces by submitting their clients' memories. Subsequently, after receiving some request of observing the performance of monitored nodes, the server-end will process the data queried from database according to some rules before responding clients with two-level monitoring data streams (compressed data level and original data level). Through carefully testing, we testify this mechanism is indeed valuable in querying and presenting large-scale monitoring data.

关键词： Large-scale systems Monitoring Testing Bandwidth Databases grid computing Delay Data security Computers Web and internet services

来源：评论

学校读者我要写书评

暂无评论

OCTOPUS: A Hybrid Scheduling Strategy for P2P VoD services

OCTOPUS: A Hybrid Scheduling Strategy for P2P VoD Services

引用

第六届网格与协同计算国际会议(The Sixth International Conference on grid and Cooperative computing GCC 2007)

作者： Xiaofei Liao Hai Jin Services Computing Technology and System Lab Cluster and Grid Computing Lab School of Computer Science and Technology Huazhong University of Science and Technology Wuhan 430074 China

Peer-to-Peer (P2P) networks have been shown to be a promising approach to provide large-scale Video on Demand (VoD) services over Internet for its potential high scalability. However, for a normal peer,how to efficiently schedule media data to multiple asynchronous peers for VoD services in such networks remains a major challenge. Hereby we propose a hybrid scheduling scheme for data dissemination in P2P VoD system in this paper, called OCTOPUS,which is combined with adaptive scheduling scheme at normal time and special scheduling strategy emergently. Usually each peer is given a value,meaning scheduling bandwidth to represent service capacity, which is an integrated evaluation for peer's bandwidth and services ability, also including statistical characteristics about historical data transferring. According to the value, one peer can adjust the expectation of each neighbor's abilities and send scheduling request of media data to all *** case of failure scheduling, emergent scheduling scheme will be called and limited data requests will be sent to media servers. OCTOPUS scheme helps to make sufficient utilization of end-peers' resource,alleviate the load of source server and improve their scalability. Experimental result from a P2P VoD system, gridCast, based on OCTOPUS schemes,shows that OCTOPUS achieves obvious effect.

关键词： Peer to peer computing Streaming media Bandwidth Network servers Processor scheduling Scalability grid computing Internet Web server Computers

来源：评论

学校读者我要写书评

暂无评论

Negotiated Coscheduling in a P2P Based HPC Platform

Negotiated Coscheduling in a P2P Based HPC Platform

引用

第六届网格与协同计算国际会议(The Sixth International Conference on grid and Cooperative computing GCC 2007)

作者： Fei Luo Hai Jin Xiaofei Liao Qin Zhang Services Computing Technology and System Lab Cluster and Grid Computing Lab School of Computer Science and Technology Huazhong University of Science and Technology Wuhan 430074 China

To promote peer-to-peer (P2P) based high performance computing (HPC), new scheduling strategies are necessarily designed to break the limitation in current systems, which only support applications with independent tasks. Therefore, we introduce a negotiated coscheduling policy, which supports applications with interactive tasks in a structured P2P based HPC platform, P2HP-2. In this policy, with the help of a project description file, a negotiated dispatching metric with two priorityentitling algorithms is proposed, which distributes tasks across computing resources. The coscheduling is further processed with two steps: spinning for dependent data and performing tasks' ***, a reliable message-event transmission method and a load balancer are developed to improve the performance. All our preliminary results show that this scheduling policy is efficient and fault-tolerant to support applications with task-dependent relationships.

关键词： Processor scheduling Distributed computing Concurrent computing Pervasive computing grid computing Internet Computer science Peer to peer computing High performance computing Application software

来源：评论

学校读者我要写书评

暂无评论

I/O Scheduling by Using Multi-network Interfaces of Virtualization system

I/O Scheduling by Using Multi-network Interfaces of Virtuali...

引用

IEEE International Conference on Trust, Security and Privacy in computing and Communications (TrustCom)

作者： Xia Xie Haiqing Zhu Wenzhi Cao Hai Jin Pingpeng Yuan Services Computing Technology and System Laboratory Cluster and Grid Computing Lab School of Computer Science and Technology Huazhong University of Science and Technology Wuhan China

With the development of virtualization technology, it can run more and more guest domains on one single physical machine. So the network I/O demand is also rapidly increasing, for these guest domains can supply more and more services or run more and more applications. This will lead to serious contention on network I/O of physical machine. Although we can use high performance NICs (network interface cards) to improve network I/O performance, one single NIC cannot get further performance improvement which depends on hardware technology. Linux bonding can achieve high bandwidth network by using multiple NICs, but it does not guarantee QoS among guest domains. Also it would terminate one guest domain's communication because of its response mechanism of ARP. Therefore, we present an approach named Virtual Bonding to achieve high bandwidth network by using multi-NICs. It aggregates multi-NICs to provide high bandwidth and high-quality network services. It also introduces load balancing mechanism between multi-NICs. So it can make full use of every NIC. Furthermore, it makes perfect support to virtualization environment.

关键词： Bonding Bridges Bandwidth Load management Servers Linux IP networks

来源：评论

学校读者我要写书评

暂无评论

Optimization of Sparse Matrix-Vector Multiplication with Variant CSR on GPUs

Optimization of Sparse Matrix-Vector Multiplication with Var...

引用

International Conference on Parallel and Distributed systems (ICPADS)

作者： Xiaowen Feng Hai Jin Ran Zheng Kan Hu Jingxiang Zeng Zhiyuan Shao Services Computing Technology and System Laboratory Cluster and Grid Computing Lab School of Computer Science and Technology Huazhong University of Science and Technology Wuhan China

Sparse Matrix-Vector multiplication (SpMV) is one of the most significant yet challenging issues in computational science area. It is a memory-bound application whose performance mostly depends on the input matrix and the underlying architecture. Many researchers have paid more attentions on exploring a variety of optimization techniques to SpMV. One of the most promising respects is how to adapt the storage format to satisfy the underlying architecture. Alterative storage formats can largely lessen memory pressure, however, the computational resources are often underutilized. Therefore, a new storage format, which is called Compressed Sparse Row with Segmented Interleave Combination (SIC), is proposed. Stemming from Compressed Sparse Row format (CSR), SIC format employs an interleave combination pattern that combines certain amount of CSR rows to form a new SIC row. In order to further improve performance, segmented processing is also brought in. According to the empirical data, we also develop an automatic SIC-based SpMV suitable for all the matrices. Experimental results show that our approach outperforms the NVIDIA CSR vector kernel, achieving up to 12.6 × speedup. It also demonstrates a comparable performance with the Hybrid format, even with the highest 2.89 × speedup.

关键词： Graphics processing unit Kernel Sparse matrices Silicon carbide Vectors Instruction sets Computer architecture

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：