学术报告
时间: 2016-11-09 发布者: 文章来源: 必威 审核人: 浏览次数: 587

时间:2016.11.10   11:00-12:00

地点:理工楼321会议室

报告人及其简历:Young Choon Lee is currently an Australian Research Council DECRA Fellow and a senior lecturer in the Department of Computing, Macquarie University. He received his PhD degree in Computer Science from The University of Sydney, Australia in 2008. His current research interests are in the areas of distributed systems and data-intensive computing.

题目:Cloud Data Centre Efficiency: Harmonizing Application Diversity and Resource Heterogeneity in Clouds

摘要:

Contemporary society relies heavily upon ICT services powered by data centres. These services run the gamut from email and social networking to banking and welfare. Data centres are everywhere; they come in all shapes and sizes ranging from several hundred servers in a rack to hundreds of thousands of servers in public cloud systems like Amazon Web Services or Google. Despite the essential requirement by many organisations including enterprises, research labs, government agencies and educational institutes, that data centres provide massive computing power and storage capacity, the majority of these organisations struggle to utilise them efficiently. Data centre utilisation is often lower than 10% due to a desire for uninterrupted service availability, typically achieved by over-provisioning. The severity of this inefficiency is tremendous and encompasses excessive capital and operating costs, and massive carbon footprint. Two main characteristics of cloud data centres that hinder effective resource sharing are the heterogeneity and dynamicity in resources and applications. In this talk, I will discuss how these two sources of inefficiency can be turned into opportunities for the improvement of cloud data centre efficiency. The idea is that resources are virtualised in a non-uniform fashion and at fine granularity for application diversity, and applications are co-located with their access to resources being considerate to each other to deal with the dynamicity of resource usage.