Speaker: Prof. Ge Yu
Title:
I/O-Efficient Big Graph Computation (Slides)
Abstract:
The proliferation of popular online social networks such as Facebook and Twitter largely depends on graph analysis over billion-node graphs. Most graph algorithms are iterative in nature and usually generate a large number of intermediate data during iterations. The growing demand for large-scale graph data analysis tasks has led to a need of highly scalable data-intensive computing platforms. In this talk we focus on how to perform graph analysis on a cluster computing environment with I/O-efficient approach, since the memory resource of a given cluster can be easily exhausted due to the drastically growing rate of real-world graph volumes and the increased intermediate data scale in proportion to the graph volume. We developed a prototype system "HybridGraph" which employs low-cost disks and/or SSDs to store graph data and perform iterative computations efficiently through optimization mechanisms including indexing graph data, managing intermediate data based on adaptive switching, and tolerating failures without using checkpoints. Also, we report extensive experiment studies, and briefly discuss some efforts in progress to enhance the functions and performance of our system.
Bio:
Ge YU is a professor of computer science, executive dean of School of Computer Science and Engineering, Northeastern University of China. His current research interests include big data management, distributed and parallel computing, data mining and machine learning, non-structural data integration, graph data and social media search. He currently serves on the editorial board of IEEE Transactions on Knowledge and Data Engineering, and has served on the program committee for many international conferences (including VLDB, ICDE, CIKM, ER, DASFAA, ADC etc). He has published more than 200 papers in refereed journals and conferences. He received his B. Tech. from Northeastern University of China in 1982, and the PhD degree in computing science from Kyushu University of Japan in 1996. He is a member of IEEE Computer Society and ACM, CCF fellow, the vice chair of the CCF Information System Technical Committee.