Policy-based distributed resource monitoring technology

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 is indicated that PULL model resource monitoring is easy to program, and easy to get cost time monitored on all computing nodes in a computing-domain. On the other hand, PUSH model has more exact monitoring period.