摘要
为了能够充分利用网络中所有节点的资源,基于multi-agent思想和技术提出了一种新的云端计算融合模型。该模型按照节点的类型将云端计算环境进行分层,并用agent作为节点行为和资源的代表,从而实现有效的资源共享和协同工作。模型充分挖掘终端节点所蕴含的可用资源,将作业分割为各种层次的任务并有序地部署到核心节点、一般服务器节点和终端节点上,以达到资源利用最有效的目标。为了能够高效、可靠地完成任务,还提出了一种新颖的基于multi-agent复合环协同管理机制,该机制特别适用于云端计算融合环境中,能够有效地增强节点贡献资源和协同工作的稳定性,并减轻云核心层管理节点的负载。最后给出了具有参考意义的基于multi-agent的云端计算融合模型实验系统的构建方法和应用示范。
In order to utilize the resources of all nodes in networks fully,a novel conjunct model of cloud client computing based on multi-agent was proposed,dividing the computing environment into three layers in accordance with the type of node and adopting agent to represent the behavior and resource of node to realize more efficient resource sharing and collaboration.The model takes full account of available resources owned by terminal nodes and deploys tasks in an orderly manner to the core nodes,servers and terminal nodes in order to maximize the efficiency of the resource utilization.To assure that tasks could be completed efficiently and reliably,a multi-agent-based collaborative multi-ring management mechanism was also presented,which is particularly suitable for the cloud client conjunct computing environment and propitious to enhance the stability of resources sharing and nodes cooperation and relieve the load of core nodes.Finally,the construction method of experimental system and application demonstration based on the model was provided for reference.
出处
《通信学报》
EI
CSCD
北大核心
2010年第10期203-211,共9页
Journal on Communications
基金
国家教育部博士点基金资助项目(20093223120001)
教育部科技发展中心网络时代的科技论文快速共享专项研究基金资助项目(2009117)
国家重点基础研究发展计划("973"计划)基金资助项目(2011CB302903)
江苏省科技支撑计划基金资助项目(BE2009158)
江苏省高校自然科学研究基金资助项目(09KJB520010)~~
关键词
分布式计算
云计算
多AGENT
任务分配
管理
distributed computing
cloud computing
multi-agent
task allocation
management
