A Market-Based Architecture for Management of Geographically
Dispersed, Replicated Web Servers
M. Karaul, Y. Korilis, and A. Orda. First International Conference on Information and
Computation Economies, page 158-165. Charleston, SC, (1998)
Abstract
Many popular Web sites employ a set of geographically dispersed,
replicated servers to address the issue of overloaded servers and
network congestion. Such distributed Web sites require allocation
mechanisms to dispatch request in a way such that any desired
load distribution can be enforced. Unlike most traditional
approaches, we propose a technique which pushes the allocation
functionality onto the client. We argue that this approach scales
well and may result in increased performance in many cases.
Building on theoretical work based on game theory, we show that
the usage of individual replicas can be affectively controlled
with cost functions even when the clients are noncooperative. We
present the design and implementation of WebSeAl, our prototype
system realizing these techniques. WebSeAl does not require any
changes to existing client and server code, conforms to HTTP
standards, and does not generate any control messages. Experiment
results indicate that WebSeAl improves performance while imposing
little overhead.
%0 Conference Paper
%1 KaraulKO98
%A Karaul, Mehmet
%A Korilis, Yannis A.
%A Orda, Ariel
%B First International Conference on Information and
Computation Economies
%C Charleston, SC
%D 1998
%K Resource Pricing, Market-Based Computing, Balancing, Theory Allocation, Game Load
%P 158-165
%T A Market-Based Architecture for Management of Geographically
Dispersed, Replicated Web Servers
%X Many popular Web sites employ a set of geographically dispersed,
replicated servers to address the issue of overloaded servers and
network congestion. Such distributed Web sites require allocation
mechanisms to dispatch request in a way such that any desired
load distribution can be enforced. Unlike most traditional
approaches, we propose a technique which pushes the allocation
functionality onto the client. We argue that this approach scales
well and may result in increased performance in many cases.
Building on theoretical work based on game theory, we show that
the usage of individual replicas can be affectively controlled
with cost functions even when the clients are noncooperative. We
present the design and implementation of WebSeAl, our prototype
system realizing these techniques. WebSeAl does not require any
changes to existing client and server code, conforms to HTTP
standards, and does not generate any control messages. Experiment
results indicate that WebSeAl improves performance while imposing
little overhead.
@inproceedings{KaraulKO98,
abstract = {Many popular Web sites employ a set of geographically dispersed,
replicated servers to address the issue of overloaded servers and
network congestion. Such distributed Web sites require allocation
mechanisms to dispatch request in a way such that any desired
load distribution can be enforced. Unlike most traditional
approaches, we propose a technique which pushes the allocation
functionality onto the client. We argue that this approach scales
well and may result in increased performance in many cases.
Building on theoretical work based on game theory, we show that
the usage of individual replicas can be affectively controlled
with cost functions even when the clients are noncooperative. We
present the design and implementation of WebSeAl, our prototype
system realizing these techniques. WebSeAl does not require any
changes to existing client and server code, conforms to HTTP
standards, and does not generate any control messages. Experiment
results indicate that WebSeAl improves performance while imposing
little overhead.},
added-at = {2006-09-13T17:44:28.000+0200},
address = {Charleston, SC},
author = {Karaul, Mehmet and Korilis, Yannis A. and Orda, Ariel},
biburl = {https://www.bibsonomy.org/bibtex/258f922771eb89deb879ac90434259f17/jmm},
booktitle = {First International Conference on Information and
Computation Economies},
description = {JMM master bibtex},
interhash = {c494ce1bb43e2e422b9ae645ee29522b},
intrahash = {58f922771eb89deb879ac90434259f17},
jmm_note = {786},
keywords = {Resource Pricing, Market-Based Computing, Balancing, Theory Allocation, Game Load},
pages = {158-165},
timestamp = {2006-09-13T17:44:28.000+0200},
title = {A Market-Based Architecture for Management of Geographically
Dispersed, Replicated Web Servers},
year = 1998
}