Catálogo de publicaciones - libros

A new generation storage system which called Object-Based Storage system (OBS) is emerging as the foundation for building massively parallel storage system. In the OBS, data files are usually stripped into multiple objects across OBS’s nodes to improve the system I/O throughput. A fundamental problem that confronts OBS is to efficiently place and locate objects in the dynamically changing environment. In this paper, we develop EOP: an efficient algorithm based on dynamic interval mapping method for object placement and lookup services. The algorithm provides immediately rebalance data objects distribution with the nodes’ addition, deletion and capability weight changing. Results from theoretical analysis, simulation experiments demonstrate the effectiveness of our EOP algorithm.

Separated grid systems are becoming the new information islands when more and more grid systems are deployed. Grid interoperation is a direction to solve that problem. This paper introduces the implementation of data interoperation between ChinaGrid and SRB. The data interoperation between them is divided into two parts: data access from SRB to ChinaGrid and from ChinaGrid to SRB. Also this paper considers the issues about performance optimization. We get a satisfied experiment result through the optimization measures.

The co-allocation architecture was developed to enable the parallel download of datasets/servers from selected replica servers, and the bandwidth performance is the main factor that affects the internet transfer between the client and the server. Therefore, it is important to reduce the difference of finished time among replica servers, and manage changeful network performance during the term of transferring as well. In this paper, we proposed an Anticipative Recursively-Adjusting Co-Allocation scheme, to adjust the workload of each selected replica server, which handles unwarned variant network performances of the selected replica servers. The algorithm is based on the previous finished rate of assigned transfer size, to anticipate that bandwidth status on next section for adjusting the workload, and further, to reduce file transfer time in a grid environment. Our approach is usefully in unstable gird environment, which reduces the wasted idle time for waiting the slowest server and decreases file transfer completion time.

In this paper, we introduce a strongly strategy-proof combinatorial auction-based grid resource allocation system, called PheonixMarket. The key advantages of PheonixMarket are that it makes the scheduling with the time-varying job value information; guarantees the combinatorial allocation of heterogeneous resources, incents users to reveal true value information of their jobs, encourages users to contribute their redundant resources and avoids exceeding resource use by the baleful users. In the performance experiments, the economic efficiency of PheonixMarket is analyzed. We then measure the price sensitivity of PheonixMarket and make the load balance experiment based on its price ’signal’. Finally, the issue of taking the funding as a form of priority is measured in the experiments.

Considering dynamic, heterogeneous and autonomous characteristics of computing resources in the computational grid environment and the advantages of economics mechanism applied to solve the problem of resource management, a sealed-bid auction method for resource allocation on computational grids is presented. Firstly, a grid service markets framework for resource allocation in the computational grid environment is described. Secondly, a sealed-bid auction mechanism is presented, where centered on users, and driven by user’s needs. Thirdly, Bayes equilibrium point and utility, strategy and efficiency in the Bayes equilibrium state are discussed. Finally, utility function-based resources allocation algorithm is presented.

Deploying the grid services with complicated dependency and status is a big challenge in service provisioning. A new model, named as is proposed in this paper. By measuring and calculating a set of metrics for dependency, the steelyard can generate the critical path of deployment action (including deploy, undeploy, and redeploy) on demand automatically for any service distributed in specific virtual organization. A practical use case, deployment of ChinaGrid Support Platform, will be demonstrated.

Web service retrieval is a very important issue for making the paradigm of service-oriented computing more practical. In this paper, we develop an automatic conceptual indexing approach of web services and investigate its application to service retrieval at the operation level. Relevant service operations are grouped into the same service concepts using a new variant -Means algorithm, and with these service concepts, web services can be indexed conceptually. A service retrieving approach, which is supported by the conceptual indexing of web services, is devised to retrieve web services in an efficient way. Experimental results show that the proposed approach greatly improves the performance of retrieving web services.

Availability of computer resources is key factor limiting use of bioinformatics analyses as a result of the growing computational demands. Grid computing provides a way to meet these requirements. But it is complicated to build a grid for users. This paper describes an approach to solve this problem using Grid Service technologies. Building the grid based on accepted standards and platforms makes the development and deployment of the grid much easier. A bioinformatics grid computing environment (BioGrid) which consists of the distributed computing application for bioinformatics is presented in this paper. Based on this environment, we propose the architecture of bioinformatics applications which is delivered using Grid Services constructed with the Globus Toolkit 4. We developed a simple program which is defined as the client-server application with grid services. It provides users an approach of grid services to impose grid resources and customize their own grid applications.

Web service technology has generated a lot interest, but its adoption rate has been slow. This paper discusses quality of service issues which is one of the contributing factors to this slow take up, and argues that nowadays policy, which requires the service providers to ensure a certain level of QoS, is not practical and will invoke high business risks. Therefore, we introduce four innovating concepts to SOA community, and propose an extending framework to ensure on-demand capacity by decoupling classical service provider into service logic provider and QoS provider. Moreover, we present a qualitative comparison in a typical web service scenario between the model proposed, classical SOA and enhanced SOA models introduced by previous publication.

Nowadays, most of the running P2P file sharing systems, such as Gnutella, generally adopt an unstructured topology and flooding search algorithms, which facing very serious search efficiency problem. In this paper, we proposed a novel intelligent link selection algorithm to address the search efficiency problem by exploiting the principle of interest-based locality. Specifically, peers continually build new connections with the others peers with same interests, this enables peers find its interest files in the nearby ones. In addition, in order to avoid adding too many connections in the overlay network which may lead the flooding search produce excessive redundant message as a consequence, we adopt a dynamic balance mechanism to delete the connection between peers which brings the least useful message. The simulation study shows our algorithm can significantly cut down the reply path lengths, achieve high search success rate with smaller search scope, and reduce the total communication cost in unstructured P2P systems.