Performance analysis of non-preemptive priority with application to cloud E-marketplaces

The increase in the level of awareness by consumers about the Cloud E-marketplaces has brought rapid growth in these competitive markets. As the markets grow, most providers are implementing different service offerings based on consumers' demand. One major challenge which has not been fully discussed in these markets is the performance impact of these offerings on consumers'. The goal of this research is to model a typical cloud marketplace and evaluate the performance impact on consumers' waiting time. To address this, we model a typical Cloud under a non-preemptive multi class discipline using Queuing theory to formulate our mathematical model and then use discrete event simulator to demonstrate a real scenario. Our evaluation was based on Non preemptive priority and non-priority discipline. Our results reveal that the unconditional average waiting time remains the same but we recorded a reduction delay which is especially profitable due to the less time in the non preemptive priority over the non priority model in four out of the five classes observed. Therefore, if the overriding requirement in the Cloud design is the reduction of the delay in some classes, then these classes should be given priority.