Old Well


The Distributed and Real-Time Systems Research Group
Department of Computer Science
The University of North Carolina at Chapel Hill



Data for the UNC HTTP Traffic Model

Line
This page makes publicly available the data for the results presented in the following two papers:

What TCP/IP Protocol Headers Can Tell Us About the Web
ACM SIGMETRICS 2001/Performance 2001, June 2001.
(Get: full citation and abstract - or - PostScript (compressed) - or - PDF copy of this paper.)

Tracing the Evolution of Web Traffic: 1995-2003
ACM MASCOTS 2003, Orlando, FL, October 2003.
(Get: full citation and abstract - or - PostScript (compressed) - or - PDF copy of this paper.)

Line
These data are made available subject to the restrictions set forth in the following copywrite notice:

Copyright Notice

Copyright 1999, 2000, 2001, and 2003 The University of North Carolina at Chapel Hill.

All rights reserved. No part of this data may be sold or distributed in any form or by any means without the prior written permission of the Department of Computer Science, University of North Carolina at Chapel Hill. Distribution and use of this data is subject to the License Agreement [incorporated in this data][set forth below]. By having, retaining or using a copy of this data, you agree to be subject to the terms of the License Agreement.

License Agreement

Permission is given to copy this file and to use them locally, as long as foregoing Copyright Notice is not removed and the data name is retained unaltered. By opening, possessing, retaining, using, or having a copy of the data, you are deemed to have agreed to the terms of this License Agreement.

The data is provided strictly on an "as is" basis without warranty of any kind. Neither the University of North Carolina at Chapel Hill, its faculty, staff or students, nor anyone else who has been involved in the creation, production or delivery of the data shall be liable for any direct, indirect, consequential or incidental damages arising out of the use or inability to use the data even if such entities or persons may be advised of the possibility of such damages.

No part of this data may be sold or distributed in any form or by any means without the prior written permission of the Department of Computer Science, University of North Carolina at Chapel Hill. Your use of the data is limited to non-commercial, not-for-profit uses and activities. To secure permission to make any other use of the data, you should contact the person named below.

Contact person:

Kevin Jeffay
Department of Computer Science
University of North Carolina at Chapel Hill
email: jeffay at cs.unc.edu
phone: 919-962-1938
fax: 919-962-1799

Line

The Empirical Distributions

The following table lists the empirical distributions of HTTP connection parameters currently available.

The format of the data files is straightforward. Each file contains the cumulative distribution function of some random variable X (e.g., HTTP response size). Column 1 is a value x of this random variable, and column 2 is Pr{X<=x}. A header is also present in each file (lines starting with #).

Model Parameter Data Set 1 Data Set 2 Data Set 3 Data Set 4
Request Sizes (in bytes) Oct. 1999 Oct. 2000 April 2001 April 2003
Response Sizes (in bytes) Oct. 1999 Oct. 2000 April 2001 April 2003
Objects Per Page Oct. 1999 Oct. 2000 April 2001 April 2003
Unique TCP Connections Per Page Oct. 1999 Oct. 2000 April 2001 April 2003
Unique Server IP Addresses Per Page Oct. 1999 Oct. 2000 April 2001 April 2003
Think Times (in milliseconds) Oct. 1999 Oct. 2000 April 2001 April 2003
Top-Level Objects Request Sizes (in bytes) Oct. 1999 Oct. 2000 April 2001 April 2003
Top-Level Objects Response Sizes (in bytes) Oct. 1999 Oct. 2000 April 2001 April 2003
Embedded Objects Request Sizes (in bytes) Oct. 1999 Oct. 2000 April 2001 April 2003
Embedded Objects Response Sizes (in bytes) Oct. 1999 Oct. 2000 April 2001 April 2003
Request Sizes For Non-Persistent Connections (in bytes) Oct. 1999 Oct. 2000 April 2001 April 2003
Response Sizes For Non-Persistent Connections (in bytes) Oct. 1999 Oct. 2000 April 2001 April 2003
Request/Response Exchanges Per Persistent Connection Oct. 1999 Oct. 2000 April 2001 April 2003
Request Sizes For Persistent Connections (in bytes) Oct. 1999 Oct. 2000 April 2001 April 2003
Response Sizes For Persistent Connections (in bytes) Oct. 1999 Oct. 2000 April 2001 April 2003
Line

- Last revised: Tue Aug 5 11:42:59 EDT 2003 by jeffay at cs.unc.edu