H$^2$—A Hybrid Heterogeneity Aware Resource Orchestrator for Cloud Platforms
Ayaz Ali Khan, Muhammad Zakarya, Rahim Khan · IEEE Systems Journal · 2019
Cloud computing services are provided through datacenters that consume significant amount of energy and contribute to global warming. Largely, resources are unused due to low service demand, in principal, wasting energy and resources. Energy could be saved through efficient resource management systems; that is of utmost importance, particularly, when service providers use various kinds of sand-boxing technologies, such as bare-metal, virtual machine and/or containers to provide quality services to users. The lack of a single resource manager creates questions on infrastructure energy efficiency and service performance. In this paper, we investigate the possibility of using a single resource manager to manage datacenter heterogeneous resources, energy, and performance (hence cost) efficiently. Our evaluation suggests that$\sim$30.47% energy could be saved at cost of only 2.14% loss in performance through using a single scheduler instead of multiple schedulers.