Congestion avoidance and control
SIGCOMM '88 Symposium proceedings on Communications architectures and protocols
Analysis of Task Assignment Policies in Scalable Distributed Web-Server Systems
IEEE Transactions on Parallel and Distributed Systems
Locality-aware request distribution in cluster-based network servers
Proceedings of the eighth international conference on Architectural support for programming languages and operating systems
Analysis of SRPT scheduling: investigating unfairness
Proceedings of the 2001 ACM SIGMETRICS international conference on Measurement and modeling of computer systems
The state of the art in locally distributed Web-server systems
ACM Computing Surveys (CSUR)
Geographic Load Balancing for Scalable Distributed Web Systems
MASCOTS '00 Proceedings of the 8th International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems
Distributed Packet Rewriting and its Application to Scalable Server Architectures
ICNP '98 Proceedings of the Sixth International Conference on Network Protocols
lbnamed: A Load Balancing Name Server in Perl
LISA '95 Proceedings of the 9th USENIX conference on System administration
Design alternatives for scalable Web server accelerators
ISPASS '00 Proceedings of the 2000 IEEE International Symposium on Performance Analysis of Systems and Software
Hi-index | 0.00 |
In this paper we present a load balancing application for HTTP traffic that uses dynamic weights. We introduce a load balancing policy based on two criteria: “process time” and “network delay”. The former describes Web servers ability to process a forthcoming request, while the latter tries to estimate network conditions. Calculation of the two criteria is periodically updated. A Weighted Round Robin algorithm was implemented using the two aforementioned metrics in order to dynamically estimate the balancing weights. We confirm that the combination of the two criteria increases sensitivity and responsiveness of the application towards network conditions and therefore the performance of the whole load balancing system. Balancing decisions should not be only “load” or “connection” dependent, but also contention dependent.