Locating nearby copies of replicated Internet servers
SIGCOMM '95 Proceedings of the conference on Applications, technologies, architectures, and protocols for computer communication
Internet Web servers: workload characterization and performance implications
IEEE/ACM Transactions on Networking (TON)
Network dispatcher: a connection router for scalable Internet services
WWW7 Proceedings of the seventh international conference on World Wide Web 7
Application-layer anycasting: a server selection architecture and use in a replicated Web service
IEEE/ACM Transactions on Networking (TON)
Dynamic Load Balancing on Web-Server Systems
IEEE Internet Computing
Enhancing the Web's Infrastructure: From Caching to Replication
IEEE Internet Computing
Dynamic Load Balancing in Geographically Distributed Heterogeneous Web Servers
ICDCS '98 Proceedings of the The 18th International Conference on Distributed Computing Systems
Using smart clients to build scalable services
ATEC '97 Proceedings of the annual conference on USENIX Annual Technical Conference
Hi-index | 0.00 |
Many popular web sites get millions of hits everyday. To service a large number of requests, clusters of fully replicated web servers are used. In such a setup, the client's request has to be directed to a cluster and then to a server within the cluster in a manner that the client receives the response in minimum time. In this paper, we propose an adaptive policy of selecting the nearest cluster for a request. Proximity is assessed by the round trip delay between the cluster and the client. An innovative idea is to measure this delay only for those clients who are sending a large number of requests. We have implemented this scheme, and using a test-bed which simulates the world wide web environment, compared the performance of the scheme with that of some existing schemes. The results indicate that the proposed scheme performs better, both in terms of average response time, as well as throughput.