Kinger Supriya, Kumar Rajesh, Sharma Anju
Department of Computer Science and Engineering, SGGS World University, Fatehgarh Sahib, Punjab, India.
School of Mathematics and Computer Applications, Thapar University, Patiala, India.
ScientificWorldJournal. 2014 Mar 11;2014:208983. doi: 10.1155/2014/208983. eCollection 2014.
Cloud computing has rapidly emerged as a widely accepted computing paradigm, but the research on Cloud computing is still at an early stage. Cloud computing provides many advanced features but it still has some shortcomings such as relatively high operating cost and environmental hazards like increasing carbon footprints. These hazards can be reduced up to some extent by efficient scheduling of Cloud resources. Working temperature on which a machine is currently running can be taken as a criterion for Virtual Machine (VM) scheduling. This paper proposes a new proactive technique that considers current and maximum threshold temperature of Server Machines (SMs) before making scheduling decisions with the help of a temperature predictor, so that maximum temperature is never reached. Different workload scenarios have been taken into consideration. The results obtained show that the proposed system is better than existing systems of VM scheduling, which does not consider current temperature of nodes before making scheduling decisions. Thus, a reduction in need of cooling systems for a Cloud environment has been obtained and validated.
云计算已迅速成为一种被广泛接受的计算范式，但对云计算的研究仍处于早期阶段。云计算提供了许多先进特性，但仍存在一些缺点，如运营成本相对较高以及存在环境危害，如碳足迹增加。通过对云资源进行高效调度，这些危害可以在一定程度上得到降低。机器当前运行的工作温度可作为虚拟机（VM）调度的一个标准。本文提出了一种新的主动式技术，该技术借助温度预测器，在做出调度决策之前考虑服务器机器（SM）的当前温度和最大阈值温度，从而使温度永远不会达到最大值。已考虑了不同的工作负载场景。所获得的结果表明，所提出的系统优于现有的虚拟机调度系统，现有系统在做出调度决策之前不考虑节点的当前温度。因此，已实现并验证了云环境中冷却系统需求的减少。
