MapReduce is emerging as an important programming model for data-intensive application. Adapting this model to desktop grid would allow taking advantage of the vast amount of computing power and distributed storage to...
详细信息
ISBN:
(纸本)9780769548159
MapReduce is emerging as an important programming model for data-intensive application. Adapting this model to desktop grid would allow taking advantage of the vast amount of computing power and distributed storage to execute new range of application able to process enormous amount of data. In 2010, we have presented the first implementation of MapReduce dedicated to Internet Desktop grid based on the BitDew middleware. In this paper, we present new optimizations to BitDew-MapReduce (BitDew-MR): aggressive task backup, intermediate result backup, task re-execution mitigation and network failure hiding. We propose a new experimental framework which emulates key fundamental aspects of Internet Desktop grid. Using the framework, we compare BitDew-MR and the open-source Hadoop middleware on grid5000. Our experimental results show that 1) BitDew-MR successfully passes all the stress-tests of the framework while Hadoop is unable to work in typical wide-area network topology which includes PC hidden behind firewall and NAT;2) BitDew-MR outperforms Hadoop performances on several aspects: scalability, fairness, resilience to node failures, and network disconnections.
MCU resource pool is the key technology of the current video conference systems. By enabling flexible and efficient allocation of MCU ports, it greatly improves the reliability of video conference system, and has been...
详细信息
ISBN:
(纸本)9781509051540
MCU resource pool is the key technology of the current video conference systems. By enabling flexible and efficient allocation of MCU ports, it greatly improves the reliability of video conference system, and has been widely used in large-scale enterprises. This paper describes the MCU resource pool scheduling strategy in State grid Corporation, analyzes the key factors of the system, and proposed a quasi-dynamic video conference resource backup strategy optimization method, which can dynamically adapt to changes in the network delay and MCU resources.
This paper proposes the seamless mode transfer technique for the parallel operation of grid-connected inverters. To meet the microgrid sources, the inverter of distributed generator is defined as a voltage source unde...
详细信息
ISBN:
(纸本)9781538660546
This paper proposes the seamless mode transfer technique for the parallel operation of grid-connected inverters. To meet the microgrid sources, the inverter of distributed generator is defined as a voltage source under the islanded mode to supply the power to the full local load demand without connection to the grid. On the other hand, if the grid is restored from its pre-fault or the strategic islanding is no more necessary, the inverter changes the operation mode to the grid-connected mode. During this mode transfer between the grid-connected and islanded modes, the load voltage should be kept constantly to meet the power quality for the continuous operation of sensitive load and to increase the load lifespan. In this paper, the PR based seamless mode transfer method is adopted with considering two following conditions. Firstly, the seamless mode transfer sequence for three-phase inverter operation is studied. Secondly, the operation technique for parallel inverters is added under the islanded mode that the load power can be supplied from the DG based parallel inverters. In this paper, the design and control of parallel inverters using the droop control based indirect current control is analyzed in detail for both islanded mode and grid-connected mode, respectively. Finally, the validity of the proposed seamless mode transfer technique is verified though the PSiM simulation results.
At the Laboratory of Information Technologies of the Joint Institute for Nuclear Research new software for job parallel calculation on the cloud resources and computing cluster is in progress. It is expected that crea...
详细信息
Smart grid becomes the hot topic concerning future power grid development trend home and abroad in recent years .In this paper the background, concept, major characteristics, core technologies are described .According...
详细信息
gridcomputing emerges as an effective technology to couple geographically distributed resources and solve large-scale computational problems in wide area networks. There are several popular grid platforms, such as UN...
详细信息
ISBN:
(纸本)9780769530482
gridcomputing emerges as an effective technology to couple geographically distributed resources and solve large-scale computational problems in wide area networks. There are several popular grid platforms, such as UNICORE and Globus Toolkit. Many famous grid projects are constructed with these platforms, each of which manages different distributed resources. In order to aggregate resources managed by different grid platforms, it is necessary to implement the interoperation among these grid platforms. Security interoperation is one of key issues for the interoperation. CGSP (Chinagrid Support Platform) and VEGA are two popular grid platforms in China. To achieve security interoperation between these two platforms, four mapping policies, including user user mapping, user-agora mapping, role-user mapping, role-agora mapping, are proposed in this paper. The security interoperation process is proposed based on SAML specification and the mapping policies. We also analyze the flexibility of mapping policies and the security of the interoperation process.
We introduce SpDISTAL, a compiler for sparse tensor algebra that targets distributed systems. SpDISTAL combines separate descriptions of tensor algebra expressions, sparse data structures, data distribution, and compu...
详细信息
This paper mainly studies the distributed PV interface device how to realize the control of distributed power grid, and the interface to achieve the function of islanding prevention, the integration of relay protectio...
详细信息
ISBN:
(纸本)9781510821934
This paper mainly studies the distributed PV interface device how to realize the control of distributed power grid, and the interface to achieve the function of islanding prevention, the integration of relay protection, measurement and control, power quality control, remote control, harmonic control functions to achieve unified management. It can be used to connect the distributed photovoltaic power generation system and the grid, to achieve unified management of distributed PV, to solve the problems of large harmonic generation in distributed generation units at low power operation without any impact on the function of power grid.
gridcomputing is now in widespread use, which integrates geographical computing resources across multiple virtual organizations to achieve high performance computing. We build a resource broker on multiple grid envir...
详细信息
ISBN:
(纸本)9783642131356
gridcomputing is now in widespread use, which integrates geographical computing resources across multiple virtual organizations to achieve high performance computing. We build a resource broker on multiple grid environments, which integrates a number of single grids from different virtual organizations without the limit of cross-organization. We can efficiently use the multiple grid resources to avoid the waste of resources. In addition, we proposed a Multi grid Resource Selection Strategy for the resource broker to select the better allocation of resources before submitting job for avoiding network congestion caused by the decrement of performance.
parallel algorithm of time-independent Monte Carlo transport problem is successful because particles are typically independent and easily distributed to multiple processors. However, for time-dependent Monte Carlo tra...
详细信息
ISBN:
(纸本)0889865701
parallel algorithm of time-independent Monte Carlo transport problem is successful because particles are typically independent and easily distributed to multiple processors. However, for time-dependent Monte Carlo transport problem, due to the communication of each time-step about scattering source attribute and meshes, it reduces the parallel efficiency and limits enlarge of parallel scale. We research two algorithms. Where adaptive processor assignment and the optimized processor choice are obtained. We propose a scheme that is based upon Monte Carlo stratified sampling technique for treatment of scattering source communication. It greatly reduces the communication cost. The parallel scalability is obviously improved. The larger speedup over the basic algorithm is obtained.
暂无评论