A Tiered System for Serving Differentiated Content |
| |
Authors: | Huamin Chen Arun Iyengar |
| |
Affiliation: | (1) Department of Computer Science, University of California at Davis, Davis, CA 95616, USA;(2) IBM T. J. Watson Research Center, P.O. Box 704, Yorktown Heights, NY 10598, USA |
| |
Abstract: | Contemporary Web sites typically consist of front–end Web servers, application servers, and back-end information systems such as database servers. There has been limited research on how to provide overload control and service differentiation for the back-end systems. In this paper we propose an architecture called tiered service (TS) for these purposes. In TS, there are several heterogeneous back-end systems to serve the Web applications. The Web applications communicate with a routing intermediary to intelligently route the queries to the appropriate back-end servers based on various policies such as client profiles and server load. In our system the back ends may store different qualities of data; lower quality data typically requires less overhead to serve. The main contributions of this paper include (i) a tiered content replication scheme that replicates tiered qualities of data on heterogeneous back ends with different capacity to satisfy clients with diverse requirements for latency and quality of data, and (ii) an application-transparent query routing architecture that automatically routes the queries to the appropriate back ends. The architecture was implemented in our test bed, and its performance was benchmarked. The experimental results demonstrate that TS offers significant performance improvement. |
| |
Keywords: | Web back-end server tiered system load balancing service differentiation |
本文献已被 SpringerLink 等数据库收录! |
|