When accessing cloud-hosted modern applications, users often suffer a significant latency due to the long geo-distance to the central cloud. Edge computing thus emerges as an alternative paradigm that can reduce this latency by deploying services close to users. In this talk, we will analyze the methodology and limitations of popular approaches for supporting AI services on geo-distributed systems along the evolution from cloud computing to edge computing. In particular, we shall discuss how to deal with different sets of challenges in distributed machine learning over heterogeneous geo-distributed systems. We shall also present our recent studies on parameter-server based framework among networked collaborative edges.
Song Guo is a Full Professor at Department of Computing, The Hong Kong Polytechnic University. His research interests are mainly in the areas of big data, cloud computing, and distributed systems with over 450 papers published in major conferences and journals. He is the recipient of the 2018 IEEE TCBD Best Conference Paper Award, the 2018 IEEE TCGCC Best Magazine Paper Award, the 2017 IEEE Systems Journal Annual Best Paper Award, and other 5 Best Paper Awards from IEEE/ACM conferences. His work was also listed in the 2016 Annual Best of Computing: Notable Books and Articles by ACM Computing Reviews. Prof. Guo was an Associate Editor of IEEE TPDS, IEEE TETC and an IEEE ComSoc Distinguished Lecturer. He is now on the editorial board of IEEE TCC, IEEE TSUSC, IEEE TGCN, IEEE Network, etc. Prof. Guo also served as General and TPC Chair for numerous IEEE conferences. He currently serves on the Board of Governors of IEEE Communications Society.