Advanced

Disturbance Rejection and Control in Web Servers

Kjaer, Martin Ansbjerg LU (2009) In PhD Theses TFRT-1086.
Abstract
An important factor for a user of web sites on the Internet

is the duration of time between the request of a web page until an

answer has been returned. If this response time is too long, the user

is likely to abandon the web site and search for other providers of

the service. To avoid this loss of users, it is important for the web

site operator to assure that users are treated sufficiently fast. On

the other hand, it is also important to minimize the effort to

optimize profit. As these objectives often are contradictory, an

acceptable target response-time that can be formulated. The resources

are allocated in a manner that ensures that long response... (More)
An important factor for a user of web sites on the Internet

is the duration of time between the request of a web page until an

answer has been returned. If this response time is too long, the user

is likely to abandon the web site and search for other providers of

the service. To avoid this loss of users, it is important for the web

site operator to assure that users are treated sufficiently fast. On

the other hand, it is also important to minimize the effort to

optimize profit. As these objectives often are contradictory, an

acceptable target response-time that can be formulated. The resources

are allocated in a manner that ensures that long response times

do not occur, while, at the same time, using as little resources as

possible to not overprovision.



The work presented in this doctoral thesis takes a

control-theoretic perspective to solve this problem. The resources

are considered as the control input, and the response time as the main

output. Several disturbances affect the system, such as the arrival

rate of requests to the web site. A testbed was designed to

allow repeatable experiments with different controller

implementations. A server was instrumented with sensors and actuators

to handle requests from 12 client computers with capability for

changing work loads.



On the theoretical side, a model of a web server is presented in this

thesis. It explicitly models a specific sensor implementation where

buffering occurs in the computer prior to the sensor. As a result, the

measurement of the arrival rate becomes state dependent under high

load. This property turns out to have some undesirable effects on the

controlled system. The model was capable of predicting the behavior of

the testbed quite well.



Based on the presented model, analysis shows that

feed-forward controllers suggested in the literature can lead to

instability under certain circumstances at high load. This has not

been reported earlier, but is in this doctoral thesis

demonstrated by both simulations and experiments. The analysis

explains why and when the instability arises.



In the attempt to predict future response-times this thesis also

presents a feedback based prediction scheme. Comparisons between

earlier predictions to the real response-times are used to correct a

model based response time prediction. The prediction scheme is

applied to a controller to compensate for disturbances before the

effect propagates to the response time. The method improves the

transient response in the case of sudden changes in the arrival

rate of requests.



This doctoral thesis also presents work on a control solution

for reserving CPU capacity for a given process or a given group of

processes on a computer system. The method uses only existing

operating-system infrastructure, and achieves the desired

CPU capacity in a soft real-time manner. (Less)
Please use this url to cite or link to this publication:
author
supervisor
opponent
  • Associate Professor Abdelzaher, Tarek F., University of Illinois at Urbana Champaign, USA
organization
publishing date
type
Thesis
publication status
published
subject
keywords
feed-forward, dynamic prediction, computer systems, queuing theory, Web server
in
PhD Theses
volume
TFRT-1086
pages
178 pages
publisher
Department of Automatic Control, Lund Institute of Technology, Lund University
defense location
Room 1406, building E, Ole Römers väg 1, Faculty of Engineering, Lund University
defense date
2009-11-20 13:15
ISSN
0280-5316
language
English
LU publication?
yes
id
e81378aa-35b0-472f-ac91-cbd87234e0ce (old id 1488247)
date added to LUP
2009-10-23 13:29:05
date last changed
2016-09-19 08:44:45
@phdthesis{e81378aa-35b0-472f-ac91-cbd87234e0ce,
  abstract     = {An important factor for a user of web sites on the Internet<br/><br>
is the duration of time between the request of a web page until an<br/><br>
answer has been returned. If this response time is too long, the user<br/><br>
is likely to abandon the web site and search for other providers of<br/><br>
the service. To avoid this loss of users, it is important for the web<br/><br>
site operator to assure that users are treated sufficiently fast. On<br/><br>
the other hand, it is also important to minimize the effort to<br/><br>
optimize profit. As these objectives often are contradictory, an<br/><br>
acceptable target response-time that can be formulated. The resources<br/><br>
are allocated in a manner that ensures that long response times<br/><br>
do not occur, while, at the same time, using as little resources as<br/><br>
possible to not overprovision.<br/><br>
<br/><br>
The work presented in this doctoral thesis takes a<br/><br>
control-theoretic perspective to solve this problem. The resources<br/><br>
are considered as the control input, and the response time as the main<br/><br>
output. Several disturbances affect the system, such as the arrival<br/><br>
rate of requests to the web site. A testbed was designed to<br/><br>
allow repeatable experiments with different controller<br/><br>
implementations. A server was instrumented with sensors and actuators<br/><br>
to handle requests from 12 client computers with capability for<br/><br>
changing work loads.<br/><br>
<br/><br>
On the theoretical side, a model of a web server is presented in this<br/><br>
thesis. It explicitly models a specific sensor implementation where<br/><br>
buffering occurs in the computer prior to the sensor. As a result, the<br/><br>
measurement of the arrival rate becomes state dependent under high<br/><br>
load. This property turns out to have some undesirable effects on the<br/><br>
controlled system. The model was capable of predicting the behavior of<br/><br>
the testbed quite well.<br/><br>
<br/><br>
Based on the presented model, analysis shows that <br/><br>
feed-forward controllers suggested in the literature can lead to<br/><br>
instability under certain circumstances at high load. This has not<br/><br>
been reported earlier, but is in this doctoral thesis<br/><br>
demonstrated by both simulations and experiments. The analysis<br/><br>
explains why and when the instability arises.<br/><br>
<br/><br>
In the attempt to predict future response-times this thesis also<br/><br>
presents a feedback based prediction scheme. Comparisons between<br/><br>
earlier predictions to the real response-times are used to correct a<br/><br>
model based response time prediction. The prediction scheme is<br/><br>
applied to a controller to compensate for disturbances before the<br/><br>
effect propagates to the response time. The method improves the<br/><br>
transient response in the case of sudden changes in the arrival<br/><br>
rate of requests.<br/><br>
<br/><br>
This doctoral thesis also presents work on a control solution<br/><br>
for reserving CPU capacity for a given process or a given group of<br/><br>
processes on a computer system. The method uses only existing<br/><br>
operating-system infrastructure, and achieves the desired<br/><br>
CPU capacity in a soft real-time manner.},
  author       = {Kjaer, Martin Ansbjerg},
  issn         = {0280-5316},
  keyword      = {feed-forward,dynamic prediction,computer systems,queuing theory,Web server},
  language     = {eng},
  pages        = {178},
  publisher    = {Department of Automatic Control, Lund Institute of Technology, Lund University},
  school       = {Lund University},
  series       = {PhD Theses},
  title        = {Disturbance Rejection and Control in Web Servers},
  volume       = {TFRT-1086},
  year         = {2009},
}