报告人：熊乃学 博士（Georgia State University）

报告题目：A Class of Failure Detection Schemes for Fault-tolerant Cloud Computing Networks

摘要：Cloud computing is an increasingly important solution for providing services deployed in dynamically scalable cloud networks. In cloud-based networks, services may be virtualized with specific server hosting details abstracted, with some servers active and available, others busy or heavily loaded, and others are offline for various reasons. Therefore, in order to provide an effective control scheme and parameter guidance for service conditions and cloud resource, failure detection is essential to meet user service expectations. Effective failure detection is a fundamental factor in addressing possible performance bottlenecks in providing virtual service. It is advisable to design and implement an acceptable failure detector (FD) for cloud-based networks. This paper first presents a general traffic-feature analysis method for optimizing existing FDs in fault-tolerant wireless networks. Based on the general method, we propose a novel FD, called Exponential Distribution Failure Detector (ED FD), which considers the probability distribution properties of arrival interval periods as exponential distribution, not traditional normal distribution. In addition, we prove that ED FD belongs to the class 3Pac (accruement property and upper bound property). Experiments have been carried out in a wireless network. The experimental results demonstrate that the proposed ED FD can outperform the existing FDs in terms of short detection time, low mistake rate, and high query accuracy probability. Finally, we present lots of significant future development along this track.

报告时间：2010年6月18日上午9：30

报告地点： 数计学院5-106

参加人员：计算机系全体教师和在新校区全体研究生

院科研办