Southwest Jiaotong University School of Mathematics


首页  >  学术科研  >  学术交流  >  正文

代数编码及其应用前沿系列讲座-法国Télécom ParisTech和CentraleSupélec大学闫起发博士学术报告

来源:   作者:代数编码及其应用团队     日期:2018-12-25 09:59:53   点击数:  

法国Télécom ParisTechCentraleSupélec






报告地点: X7503


TitleStorage, Computation, and Communication: A Fundamental Tradeoff in Distributed Computing

BiographyQifa Yan received the B.S. degree in mathematics and applied mathematics from Shanxi University, Taiyuan, China, in 2010. He obtained the Ph.D degree in communication and information system at the School of  Information Science and Technology, Southwest Jiaotong University, Chengdu, China in 2017. He is currently working as a joint Postdoctoral at Télécom Paris Tech and CentraleSupélec in France. His research interests include caching networks, distributed computing systems, and other fields related to wireless networks,  information theory and coding theory.

Abstract: Distributed computing has become one of the most important frameworks in dealing with large computation tasks. In this talk, we investigate a MapReduce like distributed computing system. We will characterize the optimal tradeoff between storage space, computation load, and communication load. The corner points of the optimal tradeoff surface are achieved by the modified coded distributed computing (M-CDC) scheme proposed by Ezzeldin et al, and time- and memory- sharing between these points achieves general surface.  We also derive an information-theoretical converse, which exactly matches the achievability. Our result thus extends the result by Li et al. on the optimum tradeoff between storage and communication to account also for the computation load. We further show how to obtain a distributed computing scheme from any placement delivery array (PDA) whose ordinary symbols occur at least twice. Previously proposed PDAs to solve the subpacketization problem in coded caching. allow us then to derive optimal distributed computing schemes that require only a small number of files, and thus have reduced complexity.