Big data applications store data sets through sharing data center under the Cloud computing environment, but the need of data set in big data applications is dynamic change over time. In face of multiple data centers,...
详细信息
ISBN:
(纸本)9781509050819
Big data applications store data sets through sharing data center under the Cloud computing environment, but the need of data set in big data applications is dynamic change over time. In face of multiple data centers, such applications meet new challenges in data migration which mainly include how to how to reduce the number of network access, how to reduce the overall time consumption, and how to improve the efficiency by the time of balancing the global load in the migration process. Facing these challenges, we first build the problem model and descript the dynamic migration method, then solve the global time consumption of data migration, the number of network access and global load balancing these three parameters. Finally, do the cloud computing simulation experiment under the Cloudsim experiment platform. The result shows that the proposed method makes the task completion time reduced by 10% and the data transmission time accounts for the proportion of the total time is reduced. When the amount of data sets is increase, the proportion can reduces to 50% or less. Network access number lower than Zipf and reached stable, in global load, the variance of the node's store space closed to zero.
The correctness of applications that perform asynchronous message passing typically relies on the underlying hardware having a sufficient amount of memory (message buffers) to hold all undelivered messages-such applic...
详细信息
ISBN:
(纸本)9780889866386
The correctness of applications that perform asynchronous message passing typically relies on the underlying hardware having a sufficient amount of memory (message buffers) to hold all undelivered messages-such applications may deadlock when executed on a system with an insufficient number of message buffers. Thus, determining the minimum number of buffers that an application needs to prevent deadlock is an important task when writing or porting parallel applications. Unfortunately, both this problem (called the Buffer Allocation Problem) and the simpler problem of determining whether an application may deadlock for a given number of available message buffers are intractable [1]. We present a new epoch-based polynomial-time approach for approximating the Buffer Allocation Problem. Our approach partitions application executions into epochs and intersperses barrier synchronizations between them, thus limiting the number of message buffers necessary to ensure deadlock-freedom. This approach produces near optimal solutions for many common cases and can be adapted to guide application modifications that ensure deadlock freedom when the application is ported. Lastly, we describe a space-time trade-off between the number of available message buffers and the number of barrier synchronizations, and describe how this trade-off can be used to fine-tune application performance.
The next generation of scientific experiments and studies, popularly called e-Science, is carried out by large collaborations of researchers distributed around the world engaged in the analysis of huge collections of ...
详细信息
The next generation of scientific experiments and studies, popularly called e-Science, is carried out by large collaborations of researchers distributed around the world engaged in the analysis of huge collections of data generated by scientific instruments. gridcomputing has emerged as an enabler for e-Science as it permits the creation of virtual organizations that bring together communities with common objectives. Within a community, data collections are stored or replicated on distributed resources to enhance storage capability or the efficiency of access. In such an environment, scientists need to have the ability to carry out their studies by transparently accessing distributed data and computational resources. In this paper, we propose and develop a grid broker that mediates access to distributed resources by: (a) discovering suitable data and computational resources sources for a given analysis scenario;(b) optimally mapping analysis jobs to resources;(c) deploying and monitoring job execution on selected resources;(d) accessing data from local or remote data sources during job execution;and (e) collating and presenting results. The broker supports a declarative and dynamic parametric programming model for creating grid applications. We have used this model in grid-enabling a high-energy physics analysis application (the Belle Analysis Software Framework). The broker has been used in deploying Belle experimental data analysis jobs on a grid testbed, called the Belle Analysis Data grid, having resources distributed across Australia interconnected through GrangeNet. Copyright (c) 2005 John Wiley & Sons, Ltd.
Tracing resource usage by grid users is of utmost importance especially in the context of large-scale scienti c collaborations such as within the High Energy Physics (HEP) community to guarantee fairness of resource s...
详细信息
ISBN:
(纸本)9780769530642
Tracing resource usage by grid users is of utmost importance especially in the context of large-scale scienti c collaborations such as within the High Energy Physics (HEP) community to guarantee fairness of resource sharing, but many dif culties can arise when tracing the resource usage of distributed applications over heterogeneous grid platforms. These dif culties are often related to a lack of interoperability of the accounting components across middlewares. This paper brie y describes the architecture and workflow of the distributedgrid Accounting System (DGAS) [1] and evaluates the possibility to extend it with a Resource Usage Service (RUS) [2, 3] interface according to the Open grid Forum (OGF) speci cation that allows to store and retrieve OGF Usage Records (URs) [4, 5] via Web Services. In this context the OGF RUS and UR speci cations are critically analyzed. Furthermore, a prototype of a RUS interface for DGAS (DGAS-RUS) is presented and the most recent test results towards a full interoperability between heterogeneous grid platforms are outlined.
Wireless grid is a collection of distributedcomputing resources such as low-powered mobile, patient monitoring sensors, wired grid and nomadic devices over a wide geographical connectivity that appear to an end-user ...
详细信息
ISBN:
(纸本)9780769533223
Wireless grid is a collection of distributedcomputing resources such as low-powered mobile, patient monitoring sensors, wired grid and nomadic devices over a wide geographical connectivity that appear to an end-user as one large virtual computing system. The wireless grid has emerged as an attractive platform to tackle various large scale problems or applications, especially in science, medical and engineering. Key services such as resource discovery, monitoring, energy monagement and scheduling are inherently more complicated in a wireless grid environment. In this paper, we present a Threshold Accepting Scheduling (TAS) mechanism to obtain the nearer-to-optimal solution faster on scheduling scientific workflows in Wireless grids. In addition, we have compared the proposed scheduling mechanism with other mechanisms, namely, simulated annealing (SA) and the combination of Genetic algorithm with Local Search (GA-LS) algorithm in terms of lateness, job completion ratio and average CPU running time. The computational results reveal that the proposed method is more effective.
grids can be considered as dominant platforms for large-scale parallel/distributedcomputing in science and engineering. Clouds allow users to acquire and release resources on-demand. Next generation computing environ...
详细信息
CMS has chosen to adopt a distributed model for all computing in order to cope with the requirements on computing and storage resources needed for the processing and analysis of the huge amount of data the experiment ...
详细信息
CMS has chosen to adopt a distributed model for all computing in order to cope with the requirements on computing and storage resources needed for the processing and analysis of the huge amount of data the experiment will be providing from LHC startup. An overview of the architecture of the CMS gridcomputing system will be given in this paper. The baseline system as well as possible extensions of the baseline capabilities and functionalities will be described. The architecture is based on a set of loosely coupled components that allow an iterative process of developing and integrating new components, increasing the scale and functionality of the system. The evolving computing system is tested in major data and service "challenges". Experience gained in past challenges and expectations from future challenges will be reported. The CMS computing environment is a distributed system of computing services and resources that interact with each other as grid services. CMS-specific services are built on top of lowerlevel services provided by grid projects. Emphasis will be placed on describing the CMS data placement and transfer system as well as the distributed Monte Carlo production on the grid. c Copyright owned by the author(s) under the terms of the Creative Commons Attribution-NonCommercial-ShareAlike Licence.
This paper introduces a fully distributed reactive power control method for a wind farm (WF) to minimize the active power losses of its internal network. Based on the modeling of wind turbines (WTs) and internal netwo...
详细信息
ISBN:
(纸本)9781538629109
This paper introduces a fully distributed reactive power control method for a wind farm (WF) to minimize the active power losses of its internal network. Based on the modeling of wind turbines (WTs) and internal network of the WF, a receding horizon optimization (RHO) based controller is proposed in this study to coordinate the var output of WTs. To calculate power losses precisely, the power flow model with second-order conic relaxation (SOCR) is presented. As the main concern, the exactness of such SOCR in a WF power flow model is proved. Then, taking advantage of the decoupled nature of branch flow model, an alternating direction method of multipliers (ADMM) based distributed solver for a WF controller is presented, by which computational burden can be alleviated through parallelcomputing and no centralized controller is needed. The effectiveness of proposed controller is demonstrated via numerical simulations of a WF.
In the last decade, different computing paradigms and modelling frameworks for the description and simulation of biochemical systems based on stochastic modelling have been proposed. From a computational point of view...
详细信息
ISBN:
(纸本)9780769549392;9781467353212
In the last decade, different computing paradigms and modelling frameworks for the description and simulation of biochemical systems based on stochastic modelling have been proposed. From a computational point of view, many simulations of the model are necessary to identify the behaviour of the system. The execution of thousands of simulation can require huge amount of time, therefore the parallelization of these algorithms is highly desirable. In this work we discuss the different strategies that can be implemented for the parallelization of a space aware tau-DPP variant, that is proving a C-MPI implementation of the system and discussing its performances according to the simulation of a particle diffusion in a crowded environment.
This paper proposes a parallel implementation of the multiple sequence alignment algorithm, known as ClustalW, on distributed memory parallel machines. The proposed algorithm divides a progressive alignment into subta...
详细信息
ISBN:
(纸本)0769523153
This paper proposes a parallel implementation of the multiple sequence alignment algorithm, known as ClustalW, on distributed memory parallel machines. The proposed algorithm divides a progressive alignment into subtasks and schedules them dynamically. A task tree is built according to the dependency of the generated phylogenetic tree. The computation and communication costs of the tasks are estimated at run-time and updated periodically. With dynamic scheduling, tasks are allocated to the processors considering the tasks' estimated computation and communication costs and the processors' workload in order to minimize the completion time. The experiment results show that the proposed parallel implementation is achieves a considerable speedup over the sequential ClustalW.
暂无评论