Share Email Print
cover

Proceedings Paper

Massive spatial data network service architecture based on double-cluster
Author(s): Zhongmin Li; Lu Gao
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

It is the tendency for the development of massive spatial data network service to use cluster to enlarge load capacity of spatial data server. In this paper, we use the OSD (Object-based Storage Device) storage cluster as the shared storage of LVS (Linux Virtual Server) server cluster, and use the servers in the server pool of the LVS server cluster as the storage client of the OSD storage cluster, to build a scalable massive spatial data network service architecture, which uses the high scalability of the LVS server cluster and the OSD storage cluster to avoid the bottlenecks of massive spatial data network service bandwidth and storage I/O throughput. Several load balance scheduling algorithms embedded in the LVS server cluster can satisfy the demand of load balance in many applications. But those algorithms can't optimize load balance of spatial data servers, regardless of the features of spatial data. Spatial data in large scale network service application is generally organized according to the global longitude and latitude, and managed according to the principle "vertical hierarchies and horizontal dividing". According to the features of spatial data, we optimize the scheduling algorithm to enhance the Cache utilization efficiency for single spatial data server.

Paper Details

Date Published: 15 October 2009
PDF: 8 pages
Proc. SPIE 7492, International Symposium on Spatial Analysis, Spatial-Temporal Data Modeling, and Data Mining, 749245 (15 October 2009); doi: 10.1117/12.837435
Show Author Affiliations
Zhongmin Li, Nanchang Hangkong Univ. (China)
Wuhan Univ. (China)
Lu Gao, Wuhan Univ. (China)


Published in SPIE Proceedings Vol. 7492:
International Symposium on Spatial Analysis, Spatial-Temporal Data Modeling, and Data Mining
Yaolin Liu; Xinming Tang, Editor(s)

© SPIE. Terms of Use
Back to Top