Using a vision cognitive algorithm to schedule virtual machines
Jiaqi Zhao,Yousri Mhedheb,Jie Tao,Foued Jrad,Qinghuai Liu,Achim Streit
DOI: https://doi.org/10.2478/amcs-2014-0039
IF: 2.157
2014-09-01
International Journal of Applied Mathematics and Computer Science
Abstract:Abstract Scheduling virtual machines is a major research topic for cloud computing, because it directly influences the performance, the operation cost and the quality of services. A large cloud center is normally equipped with several hundred thousand physical machines. The mission of the scheduler is to select the best one to host a virtual machine. This is an NPhard global optimization problem with grand challenges for researchers. This work studies the Virtual Machine (VM) scheduling problem on the cloud. Our primary concern with VM scheduling is the energy consumption, because the largest part of a cloud center operation cost goes to the kilowatts used. We designed a scheduling algorithm that allocates an incoming virtual machine instance on the host machine, which results in the lowest energy consumption of the entire system. More specifically, we developed a new algorithm, called vision cognition, to solve the global optimization problem. This algorithm is inspired by the observation of how human eyes see directly the smallest/largest item without comparing them pairwisely. We theoretically proved that the algorithm works correctly and converges fast. Practically, we validated the novel algorithm, together with the scheduling concept, using a simulation approach. The adopted cloud simulator models different cloud infrastructures with various properties and detailed runtime information that can usually not be acquired from real clouds. The experimental results demonstrate the benefit of our approach in terms of reducing the cloud center energy consumption
automation & control systems,computer science, artificial intelligence,mathematics, applied