Design and Performance Optimization of Metadata Server in Mass Storage System

(Abstract )This paper designs and implements a metadata server of GRASS mass storage system, analyzes and optimizes three performance factors of the metadata service: Metadata organization, communication performance and search efficiency, and introduces the implementation of Bloom Filter algorithm in name space search. Experimental results show that, performance of the target system is significantly improved, and it basically fulfills the requirement for large2scale computation of high energy physics.

[1]  Thomas E. Anderson,et al.  A Comparison of File System Workloads , 2000, USENIX Annual Technical Conference, General Track.

[2]  Giuseppe Lo Presti,et al.  CASTOR: A Distributed Storage Resource Facility for High Performance Data Processing at CERN , 2007, 24th IEEE Conference on Mass Storage Systems and Technologies (MSST 2007).

[3]  Burton H. Bloom,et al.  Space/time trade-offs in hash coding with allowable errors , 1970, CACM.