Performance and dependability evaluation of scalable massively parallel computer systems with conjoint simulation

摘要: Computer systems are becoming more and a part of our daily life; business industry rely on their service, the health human beings depends correct functioning. used for critical tasks have to be carefully designed tested during early design stage, prototype phase, operational life. Methods tools required support facilitate this vital task. In article, we tackle issue system-level performance dependability analysis fault-tolerant scalable computer systems. A modeling methodology called “Conjoint Simulation” is presented, which based parti tioning system model combination several techniques. Object-oriented construction process-based simulation applied architecture workload modeling, timed Petri nets core technique representing failure scenarios repair policies. Splitting overall exploiting appropriate techniques ease development, maintenance, extensibility large-scale complex models. Furthermore, provided hierarchical construction, object-oriented simulated error injection in order perform combined analysis.