基于策略的分布式资源监测技术

Policy-based Distributed Resource Monitoring Technology

资源监测技术对于分布式计算环境下进行任务分配、负载平衡、容错都具有非常重要的作用,给出了分布式环境中资源监测技术的一种资源监测构件设计思想,在策略上包括PULL和PUSH两种资源监测模式,并在基于计算域的分布适应性资源管理支持系统(ARDRM)中进行了试验分析,表明PULL模式的资源监测模式易于编程实现,可灵活地获取在一个计算域内监测所有计算节点所必须花费的时间。而PUSH模式的资源监测模式具有更精确的监测周期。

In distributed computing environment, resource monitoring is important to task allocation, load balancing and fault tolerance. Policy-base resource monitoring is presented, in strategy, it includes two resource monitoring model, PULL and PUSH. By the experimentation analysis in computing-domain-based distributed adaptive resource management system, it indicates that PULL model resource monitoring is easy to program, and easy to get cost time monitored all computing node in a computing-domain, On the other hand, PUSH model has more exact monitoring period.