Visit ShanghaiTech University | 中文 | How to find us
HOME > News and Events > Events
Active Data Reconstructions for Erasure-coded Storage Clusters
Date: 2018/7/11             Browse: 54

Speaker:     Dr. Xiao Qin. Auburn

Time:          10:00—11:00, July 11

Location:    Room 1A-200, SIST Building

Host:          Prof. Shu Yin


A key design goal of erasure-coded storage clusters is to minimize reconstruction time, which in turn leads to high reliability by reducing vulnerability window size. PULL-Rep and PULL-Sur are two existing reconstruction schemes based on PULL-type transmission. In this talk, we incorporate PUSH-type transmissions to node reconstruction, where each surviving node partially generates an intermediate block forwarded to a subsequent surviving node. We propose two PUSH-based reconstruction schemes (i.e., PUSH-Rep and PUSH-Sur), which cannot only exploit the I/O parallelism of PULL-Sur, but also maintain sequential I/O accesses inherited from PULL-Rep. In contrast to the PULL-based schemes where surviving blocks are fetched in the synchronized many-to-one traffic pattern, our PUSH-based reconstruction supports the one-to-one pattern, which naturally solves the TCP Incast problem. We articulate four models to study the reconstruction process and estimate the reconstruction time of the four schemes in large-scale storage clusters. We illustrate how to develop a proof-of-concept prototype to quantitatively evaluate the four reconstruction schemes. We show that the PUSH-based reconstruction schemes outperform the PULL-based counterparts. In a real-world (9,6)RS-coded storage cluster, PUSH-Rep speeds up the reconstruction time by a factor of 5.76 compared with PULL-Rep; PUSH-Sur accelerates the reconstruction by a factor of 1.85 relative to PULL-Sur.


Xiao Qin received the BS and the MS degrees in computer science from the Huazhong University of Science and Technology, China, and the PhD degree in computer science from the University of Nebraska-Lincoln, Lincoln, in 1992, 1999, and 2004, respectively. Currently, he is a professor in the Department of Computer Science and Software Engineering, Auburn University. His research interests include parallel and distributed systems, storage systems, fault tolerance, real-time systems, and performance evaluation. He received the U.S. NSF Computing Processes and Artifacts Award and the NSF Computer System Research Award in 2007 and the NSF CAREER Award in 2009. He is a senior member of the IEEE.

SIST-Seminar 18053