The problem of stochastically allocating redundant com- ponents to increase the system lifetime is an important topic of reliability. An optimal redundancy allocation is proposed, which maximizes the expected lifetime...The problem of stochastically allocating redundant com- ponents to increase the system lifetime is an important topic of reliability. An optimal redundancy allocation is proposed, which maximizes the expected lifetime of a reliability system with sub- systems consisting of components in parallel. The constraints are minimizing the total resources and the sizes of subsystems. In this system, each switching is independent with each other and works with probability p. Two optimization problems are studied by an incremental algorithm and dynamic programming technique respectively. The incremental algorithm proposed could obtain an approximate optimal solution, and the dynamic programming method could generate the optimal solution,展开更多
In distributed parallel server system, location and redundancy of repficas have great influence on availability and efficiency of the system. In order to improve availability and efficiency of the system, two phase de...In distributed parallel server system, location and redundancy of repficas have great influence on availability and efficiency of the system. In order to improve availability and efficiency of the system, two phase decision algorithm of replica allocation is proposed. The algorithm which makes use of auto-regression model dynamically predicts the future count of READ and WRITE operation, and then determines location and redundancy of replicas by considering availability, CPU and bands of the network. The algorithm can not only ensure the requirement of availability, but also reduce the system resources consumed by all the operations in a great scale. Analysis and test show that communication complexity and time complexity of the algorithm satisfy O（n）, resource optimizing scale increases with the increase of READ count.展开更多
随着云计算技术的发展与成熟,并行计算在云环境中得到了越来越多的实践。服务器无感知计算作为云计算中的一种新型的应用部署与计算方式,允许用户弹性分配资源并实现负载均衡,并提供了更强的可扩展性和更大的灵活性。然而,无状态的特性...随着云计算技术的发展与成熟,并行计算在云环境中得到了越来越多的实践。服务器无感知计算作为云计算中的一种新型的应用部署与计算方式,允许用户弹性分配资源并实现负载均衡,并提供了更强的可扩展性和更大的灵活性。然而,无状态的特性导致服务器无感知计算框架并不完全适用于传统并行计算,其中通信是一个关键问题。本文提出了一个具有服务质量(quality of service,QoS)保障的通信框架FreeParallel,旨在基于服务器无感知计算中的函数即服务(function as a service,FaaS)平台构建面向并行计算的通信能力。FreeParallel结合了消息传递接口(message passing interface,MPI)并行计算编程模型,有效地保证了通信服务的质量;并采用代理模型来支持并行函数的识别和转换,并以服务形式灵活部署在多个FaaS或虚拟化平台上。此外,本研究还提出了函数间通信流量的QoS管理策略fm Clock,在保证传输公平性的前提下,实现基于请求和限制的通信原语级网络资源分配。实验结果表明,点对点通信场景下FreeParallel与虚拟化平台的覆盖网络相比传输性能略有不足,但比当前服务器无感知计算状态共享方案的传输效率有至少89.5%的提升。并且FreeParallel在集合通信场景下表现极佳,比基线方法提升了59.9%~83.1%。同时,带有fm Clock策略的FreeParallel能够实现原语级按比例分配策略,避免了不同原语间请求的交叉干扰,案例表明,策略的加入降低了应用25.0%的完成时间。展开更多
To cope with the task scheduling problem under multi-task and transportation consideration in large-scale service oriented manufacturing systems(SOMS), a service allocation optimization mathematical model was establis...To cope with the task scheduling problem under multi-task and transportation consideration in large-scale service oriented manufacturing systems(SOMS), a service allocation optimization mathematical model was established, and then a hybrid discrete particle swarm optimization-genetic algorithm(HDPSOGA) was proposed. In SOMS, each resource involved in the whole life cycle of a product, whether it is provided by a piece of software or a hardware device, is encapsulated into a service. So, the transportation during production of a task should be taken into account because the hard-services selected are possibly provided by various providers in different areas. In the service allocation optimization mathematical model, multi-task and transportation were considered simultaneously. In the proposed HDPSOGA algorithm, integer coding method was applied to establish the mapping between the particle location matrix and the service allocation scheme. The position updating process was performed according to the cognition part, the social part, and the previous velocity and position while introducing the crossover and mutation idea of genetic algorithm to fit the discrete space. Finally, related simulation experiments were carried out to compare with other two previous algorithms. The results indicate the effectiveness and efficiency of the proposed hybrid algorithm.展开更多
基金Manuscript received March 5, 2010 accepted March 2, 2011 Supported by National Natural Science Foundation of China （61004103）, National Research Foundation for the Doctoral Program of Higher Education of China （20100111110005）, China Postdoctoral Science Foundation （20090460742）, and Natural Science Foundation of Anhui Province of China （090412058, 11040606Q44）
基金supported by the National Natural Science Foundation of China(7117217271101158+3 种基金71272058)the Program for New Century Excellent Talents in University(NCET-10-0043)the Key Project Cultivation Fund of the Scientific and Technical Innovation Program of Beijing Institute of Technology(2011CX01001)the Special Fund of International Science and Technology Cooperation Program of Beijing Institute of Technology(GZ2014215101)
文摘The problem of stochastically allocating redundant com- ponents to increase the system lifetime is an important topic of reliability. An optimal redundancy allocation is proposed, which maximizes the expected lifetime of a reliability system with sub- systems consisting of components in parallel. The constraints are minimizing the total resources and the sizes of subsystems. In this system, each switching is independent with each other and works with probability p. Two optimization problems are studied by an incremental algorithm and dynamic programming technique respectively. The incremental algorithm proposed could obtain an approximate optimal solution, and the dynamic programming method could generate the optimal solution,
文摘In distributed parallel server system, location and redundancy of repficas have great influence on availability and efficiency of the system. In order to improve availability and efficiency of the system, two phase decision algorithm of replica allocation is proposed. The algorithm which makes use of auto-regression model dynamically predicts the future count of READ and WRITE operation, and then determines location and redundancy of replicas by considering availability, CPU and bands of the network. The algorithm can not only ensure the requirement of availability, but also reduce the system resources consumed by all the operations in a great scale. Analysis and test show that communication complexity and time complexity of the algorithm satisfy O（n）, resource optimizing scale increases with the increase of READ count.
文摘随着云计算技术的发展与成熟,并行计算在云环境中得到了越来越多的实践。服务器无感知计算作为云计算中的一种新型的应用部署与计算方式,允许用户弹性分配资源并实现负载均衡,并提供了更强的可扩展性和更大的灵活性。然而,无状态的特性导致服务器无感知计算框架并不完全适用于传统并行计算,其中通信是一个关键问题。本文提出了一个具有服务质量(quality of service,QoS)保障的通信框架FreeParallel,旨在基于服务器无感知计算中的函数即服务(function as a service,FaaS)平台构建面向并行计算的通信能力。FreeParallel结合了消息传递接口(message passing interface,MPI)并行计算编程模型,有效地保证了通信服务的质量;并采用代理模型来支持并行函数的识别和转换,并以服务形式灵活部署在多个FaaS或虚拟化平台上。此外,本研究还提出了函数间通信流量的QoS管理策略fm Clock,在保证传输公平性的前提下,实现基于请求和限制的通信原语级网络资源分配。实验结果表明,点对点通信场景下FreeParallel与虚拟化平台的覆盖网络相比传输性能略有不足,但比当前服务器无感知计算状态共享方案的传输效率有至少89.5%的提升。并且FreeParallel在集合通信场景下表现极佳,比基线方法提升了59.9%~83.1%。同时,带有fm Clock策略的FreeParallel能够实现原语级按比例分配策略,避免了不同原语间请求的交叉干扰,案例表明,策略的加入降低了应用25.0%的完成时间。
基金Project(2012B091100444)supported by the Production,Education and Research Cooperative Program of Guangdong Province and Ministry of Education,ChinaProject(2013ZM0091)supported by Fundamental Research Funds for the Central Universities of China
文摘To cope with the task scheduling problem under multi-task and transportation consideration in large-scale service oriented manufacturing systems(SOMS), a service allocation optimization mathematical model was established, and then a hybrid discrete particle swarm optimization-genetic algorithm(HDPSOGA) was proposed. In SOMS, each resource involved in the whole life cycle of a product, whether it is provided by a piece of software or a hardware device, is encapsulated into a service. So, the transportation during production of a task should be taken into account because the hard-services selected are possibly provided by various providers in different areas. In the service allocation optimization mathematical model, multi-task and transportation were considered simultaneously. In the proposed HDPSOGA algorithm, integer coding method was applied to establish the mapping between the particle location matrix and the service allocation scheme. The position updating process was performed according to the cognition part, the social part, and the previous velocity and position while introducing the crossover and mutation idea of genetic algorithm to fit the discrete space. Finally, related simulation experiments were carried out to compare with other two previous algorithms. The results indicate the effectiveness and efficiency of the proposed hybrid algorithm.
