Unexpected loads in Cloud data centers may trigger overloaded situation andperformance degradation. To guarantee system performance, cloud computingenvironment is required to have the ability to handle overloads. The existingapproaches, like Dynamic Voltage Frequency Scaling and VM consolidation, areeffective in handling partial overloads, however, they cannot function when thewhole data center is overloaded. Brownout has been proved to be a promisingapproach to relieve the overloads through deactivating applicationnon-mandatory components or microservices temporarily. Moreover, brownout hasbeen applied to reduce data center energy consumption. It shows that there aretrade-offs between energy saving and discount offered to users (revenue loss)when one or more services are not provided temporarily. In this paper, wepropose a brownout-based approximate Markov Decision Process approach toimprove the aforementioned trade-offs. The results based on real tracedemonstrate that our approach saves 20% energy consumption than VMconsolidation approach. Compared with existing energy-efficient brownoutapproach, our approach reduces the discount amount given to users while savingsimilar energy consumption.
展开▼
