It was investigational on the basis of the considered existent methods, that the task of exposure of optimal recovery scheme answers the mathematical task of BIB-schemes. In the process of work the recovery scheme was...
详细信息
ISBN:
(纸本)9781509037377
It was investigational on the basis of the considered existent methods, that the task of exposure of optimal recovery scheme answers the mathematical task of BIB-schemes. In the process of work the recovery scheme was worked out on the basis of BIB-schemes, which works for the greater number of the damaged computers, what other existent charts. Recovery scheme of distributed system that guarantees optimal load of distribution in the worst case, when k computers went down, is an optimal recovery scheme with criteria of even load value of n and k based on IRB. This recovery scheme to provide optimal renewal for any amount of computers in a cluster.
distributed cloud has been widely adopted to support service requests from dispersed regions, especially for large enterprise which requests virtual desktops for multiple geodistributed branch companies. The cloud ser...
详细信息
distributed cloud has been widely adopted to support service requests from dispersed regions, especially for large enterprise which requests virtual desktops for multiple geodistributed branch companies. The cloud service provider (CSP) aims to deliver satisfactory services at the least cost. CSP selects proper data centers (DCs) closer to the branch companies so as to shorten the response time to user request. At the same time, it also strives to cut cost considering both DC level and server level. At DC level, the expensive long distance inter-DC bandwidth consumption should be reduced and lower electricity price is sought. Inside each tree-like DC, servers are trying to be used as little as possible so as to save equipment cost and power. Innature, there is a noncooperative relation between the DC level and server level in the selection. To attain these objectives and capture the noncooperative relation, multiobjective bilevel programming is used to formulate the problem. Then a unified genetic algorithm is proposed to solve the problem which realizes the selection of DC and server simultaneously. The extensive simulation shows that the proposed algorithm outperforms baseline algorithm in both quality of service guaranteeing and cost saving.
This paper investigates the problem of allocating parallel application tasks to processors in heterogeneous distributed computing systems. The authors are proposing a simple, fast and effective algorithm to find the b...
详细信息
This paper investigates the problem of allocating parallel application tasks to processors in heterogeneous distributed computing systems. The authors are proposing a simple, fast and effective algorithm to find the best possible solution with low computation time. The proposed algorithm uses the Correlation Density Rank (CDR) algorithm on the Virtual Allocation Network (VAN) in order to obtain reliability-based influence coefficients. Those coefficients are used to construct a simple linear programing as a new framework. We study the performance of the proposed algorithm over a wide range of parameters including problem size, the ratio of average communication time to average computation time, and task interaction density. The applicability and effectiveness of our algorithm is demonstrated by comparing with recently published related algorithms, which are found in the literature.
We consider one applied global optimization problem where a set of feasible solutions is discrete and very large. The goal is to find optimal perfect gratings, which can guarantee high quality and security of the visu...
详细信息
We consider one applied global optimization problem where a set of feasible solutions is discrete and very large. The goal is to find optimal perfect gratings, which can guarantee high quality and security of the visual cryptography method. A priori estimation techniques, such as branch and bound type methods, cannot be applied to exclude an essential part of elements from the feasible set. Thus, a full search is required to solve this global optimization problem exactly, which is very computationally demanding. A library of C++ templates is developed that allows its user to implement parallel master-slave algorithms for his/her application without any knowledge of parallel programming API (application programming interface). Design of the templates allows users to build a parallel solver using MPI (message passing interface) API or distributed computing application using BOINC (Berkeley open infrastructure for network computing) API from the same C/C++ code with implementation of application-specific tasks. We build parallel and distributed computing solvers for the considered optimization problem and present results of computational experiments using a computer cluster and BOINC project for volunteer computing. Heuristic methods are also considered as an alternative to the full search algorithm. Due to complicated conditions defining feasible solutions (perfect gratings), genetic algorithms cannot be used to solve this problem efficiently. We propose two memetic heuristic algorithms, when a basic stochastic or simplified full search algorithm is combined with a local search algorithm. Parallel heuristic algorithms are also proposed and implemented. The efficiency and accuracy of heuristics are investigated and results of experiments are presented.
This tutorial presents concepts and basics of distributed computing which are important (at least from the author's point of view!), and should be known and mastered by Master students, researchers, and engineers....
详细信息
This tutorial presents concepts and basics of distributed computing which are important (at least from the author's point of view!), and should be known and mastered by Master students, researchers, and engineers. Those include: (a) a characterization of distributed computing (which is too much often confused with parallel computing); (b) the notion of a synchronous system and its associated notions of a local algorithm and message adversaries; (c) the notion of an asynchronous shared memory system and its associated notions of universality and progress conditions; and (d) the notion of an asynchronous messagepassing system with its associated broadcast and agreement abstractions, its impossibility results, and approaches to circumvent them. Hence, the tutorial can be seen as a guided tour to key elements that constitute basics of distributed computing.
The problem of distributed function computation for the class of smooth sources is studied, where functions to be computed are compositions of symbol-wise functions and some outer functions that are not symbol-wise. T...
详细信息
The problem of distributed function computation for the class of smooth sources is studied, where functions to be computed are compositions of symbol-wise functions and some outer functions that are not symbol-wise. The optimal rate for computing those functions is characterized in terms of the Slepian-Wolf rate and an equivalence class of sources induced by functions. To prove the result, a new method to derive a converse bound for distributed computing is proposed; the bound is derived by identifying a source that is inevitably conveyed to the decoder and by explicitly constructing a code for reproducing that source. As a byproduct, it provides a conceptually simple proof of the known fact that computing a Boolean function may require as large rate as reproducing the entire source.
Unpredictable fluctuations in resource availability often lead to rescheduling decisions that sacrifice a success rate of job completion in batch job scheduling. To overcome this limitation, we consider the problem of...
详细信息
Unpredictable fluctuations in resource availability often lead to rescheduling decisions that sacrifice a success rate of job completion in batch job scheduling. To overcome this limitation, we consider the problem of assigning a set of sequential batch jobs with demands to a set of resources with constraints such as heterogeneous rescheduling policies and capabilities. The ultimate goal is to find an optimal allocation such that performance benefits in terms of makespan and utilization are maximized according to the principle of Pareto optimality, while maintaining the job failure rate close to an acceptably low bound. To this end, we formulate a multihybrid policy decision problem (MPDP) on the primary-backup fault tolerance model and theoretically show its NP-completeness. The main contribution is to prove that our multihybrid job scheduling (MJS) scheme confidently guarantees the fault-tolerant performance by adaptively combining jobs and resources with different rescheduling policies in MPDP. Furthermore, we demonstrate that the proposed MJS scheme outperforms the five rescheduling heuristics in solution quality, searching adaptability and time efficiency by conducting a set of extensive simulations under various scheduling conditions. (C) 2015 Elsevier B.V. All rights reserved.
Message exchange is a central activity in distributed computing frameworks. Nevertheless, past research has paid little attention on profiling techniques and tools for endpoint communication. In this paper, we fill th...
详细信息
Message exchange is a central activity in distributed computing frameworks. Nevertheless, past research has paid little attention on profiling techniques and tools for endpoint communication. In this paper, we fill this gap by introducing a new fine-grained profiler for endpoints and communication between them in distributed systems. Our tool aids efficient analysis, optimization, and tuning of endpoint communication in distributed frameworks. The paper presents an overview of our profiler, discusses employed profiling techniques and collected metrics, and shows preliminary results in profiling a well-known distributed computing framework.
It is investigated, on the basis of the considered existent methods, that the task of exposure of optimal recovery scheme answers the mathematical task of idea ring bundle. In the process of work the recovery scheme w...
详细信息
ISBN:
(纸本)9781509014149
It is investigated, on the basis of the considered existent methods, that the task of exposure of optimal recovery scheme answers the mathematical task of idea ring bundle. In the process of work the recovery scheme was worked out on the basis of idea ring bundle, which works for the greater number of the damaged computers, what other existent charts. This recovery scheme provides optimal renewal for any amount of computers in a cluster.
The aim of the study is to develop tools of increasing a problem solving reliability in a heterogeneous distributed computing environment by applying a diagnostics of computing resources components and using an analys...
详细信息
The aim of the study is to develop tools of increasing a problem solving reliability in a heterogeneous distributed computing environment by applying a diagnostics of computing resources components and using an analysis of problem solving schemes. A scheme (a plan) is an abstract program for a problem solving. A special attention is paid to the calculation of a problem solving scheme reliability on the basis of a logical-probabilistic method. This method is based on transiting from Boolean functions for a reliability description of a problem solving scheme to probability functions for determining indicators of such reliability. Improving a problem solving scheme reliability is carried out by a resource reservation. The resource reservation applied in a problem solving scheme provides obtaining of a reliability indicator that approximates maximally the predetermined criterion of reliability, taking into account limitations on a number of reserve resources. The example of the problem solving scheme and calculating its reliability is represented.
暂无评论