WWW usage

Tony Sanders <sanders@bsdi.com>
Errors-To: sanders@bsdi.com
Errors-To: sanders@bsdi.com
Message-id: <9306260423.AA24108@austin.BSDI.COM>
To: www-talk@nxoc01.cern.ch
Subject: WWW usage
Errors-To: sanders@bsdi.com
Reply-To: sanders@bsdi.com
Organization: Berkeley Software Design, Inc.
Date: Fri, 25 Jun 1993 23:23:39 -0500
From: Tony Sanders <sanders@bsdi.com>
FYI: This document_ is available via the setext -> HTML converter:

WWW Usage Tracking
  Is anyone interested in tracking WWW usage?

  I propose someone setup a mail gateway for usage reports that get dumped
  into the Web.  I would be willing to do this if no one else is interested.
  I could at least archive the raw data onto the Web if not summerize it and
  graph it.

  Maybe a weekly report that says something like:

    Date: Sat Jun 26 03:01:56 GMT 1993
    Server: http://www.bsdi.com/
    Requests last 7 days: 3589
    Unique hosts last 7 days: 69
    Total unique hosts: 847
    Total requests: 18358
    Total days of service: 52

  Whatdaya think?

  At some point it would be neat if someone could take all the IP addresses
  from all the servers and ``uniq'' them to get an idea of how many users
  we have (My server has 847 unique ip addresses in the log).  Does anyone
  care about having their IP address uniqued?  Does anyone care about having
  their IP address made public (I'm not suggesting this, just curious).

  Once HTTP/1.0 gets more widespread we will be able to track client usage.

  In case anyone else is interested: my 7 day summaries for the past
  52 days (minus this week).  Each * is 100 requests.

    05/03 ********************
    05/10 ********************
    05/17 ******************
    05/24 ***************
    05/31 ***********************
    06/07 *****************************
    06/14 *******************************

  Anyone else have this data they can post?

A Related Issue

  Is anyone having any performance problems with their HTTP servers?  I did
  some tests on my 486DX2/66 (call it 31MIPS) and found that Plexus runs
  about 5 tps, and I timed the NCSA server that runs from inetd but it's
  not a fair comparison (1 tps) because you could run it native and save
  the exec cost.  What are you busy server people running at peak (in tps)?
  I would **really** like to hear what tps other people are getting and on
  what hardware (~ MIPS anyway) and if you can do some peak performance
  tests that would help.  I expect it'll be about the same (5 tps peak) on
  most hardware available today as I think most of the overhead is beyond
  the control of the server (i.e., in TCP/IP and exec).


.. _document http://www.bsdi.com/setext/www-usage.etx