中文版  |  English Version

学术讲座通知:Back to the Future: Challenges in Cloud-Based Service Computing

发布时间:2017-07-10 10:40  浏览:425次

      应网络与交换技术国家重点实验室程渤教授的邀请,美国佐治亚理工学院Calton Pu教授将于7月12日来北京邮电大学作学术报告。欢迎校内广大师生踊跃参加。

讲座题目:Back to the Future: Challenges in Cloud-Based Service Computing

主讲人:Calton Pu教授

主持人:程渤

时间:2017712日(星期三)下午14:0015:00

地点:新科研楼602会议室

Abstract

Computing clouds offer information technology infrastructure as services with promises of economy of scale and low start-up costs. However, the management of modern N-tier applications contains significant challenges due to several factors that include growing and evolving complexity of applications, non-stationary workloads, virtualization and consolidated environments, and dynamic service composition. Furthermore, paramount to mission-critical service applications are extra-functional quality of service (QoS) properties such as predictable performance, continuous availability, end-to-end dependability, power consumption, privacy and security. These challenges far exceed the capability of current analytical models for capacity planning and system management. As a starting point towards cloud-based service computing, we describe an observational large-scale performance study of N-tier application benchmarks in the Elba project. More generally, we advocate an experimental approach to help N-tier applications achieve desired QoS, for example, performance goals specified in service level agreements. Meeting the challenges of QoS is an important and necessary step in the co-evolution of both Cloud Computing and Service Computing.

Bio

Calton's research interests are in the areas of service computing, distributed and cloud computing, integration and veracity of big data. His current projects include cloud computing (Elba) and big data (GRAIT-DM) research. Using experimental data from realistic benchmarks, the Elba project studies the interesting phenomena such as very short bottlenecks that have large impact on n-tier system response time. The GRAIT-DM project collects real world data from social sensors (e.g., Twitter and YouTube) and physical sensors (e.g., USGS GSN and NASA TRMM) to detect physical events and manage real-time information on them. The sponsors for Calton Pu's research include both government funding agencies such as NSF, and companies from industry such as HP, Fujitsu, IBM, and Intel.  He is a co-director of Center for Experimental Research in Computer Systems (CERCS), and affiliate faculty of Institute for Information Security and Privacy (IISP) at Georgia Tech.  He is also the director of RCN on Big Data for Smart Cities, with collaborations around the world.

 

网络与交换技术国家重点实验室

2017年7月10 日