SourceForge has been redesigned. Learn more.
Close

#29 SARG and time computation algorithm

closed
nobody
None
5
2012-08-06
2012-06-15
RoFz
No

Hello,

how does SARG compute the total amount of time users spent on a specific web page? I would like to get a better understanding on how this time calculation mechanism works since i'm getting users with spent times such as this:

www.facebook.com:443 242.58M 3.94% 0.00% 100.00% 203:46:09 14.29%

Would that '203:46:09' represent HOURS:MINUTES:SECONDS?

So, the statistic above could be interpreted as if the user spent 203 hours using Facebook? That is definitively not true.

I'll appreciate any help or suggestion of documentation.

Thanks!

Discussion

  • Frederic Marchal

    RoFz, sorry for not responding sooner. I missed your post and just got my attention attracted to it by rajawaji's response.

    The elapsed time is not the time the user spend being idle at work surfing the web. Squid (and sarg) cannot possibly know what the user is doing with the downloaded data on his/her computer. All we know is the time the proxy spent processing the data. So that's what the elapsed time is. It is the sum of all the times of all the connections to www.facebook.com. Sarg doesn't attempt to be clever. It doesn't assume some pattern in the user's behavior. We cannot know if the user read the document for three minutes, discarded it after a few seconds or played a flash game for hours.

    In your example, the proxy likely spent a total of 203 hours processing requests directed at facebook. It is a raw result without additional assumption or computation trick. Now, from my experience, it is very likely that the bulk of that time is made of unsuccessful connections that are kept open until the timeout elapses. Those connections take no bandwidth nor keep they the processor busy. They should have no consequences on your proxy.

    Now, you may still think that 203 hours is still incredibly high for a 8 hours work day (or even maybe for 24h). Don't forget that a single page downloads lots of additional files such as css, javascript, images, and so on from multiple sites (including tracking sites, advertisement sites and links to every social network such as facebook). They all add up to the elapsed time although they occurs in parallel. Users may open multiple tabs or windows on one or several sites simultaneously. For instance, I have a habit of opening multiple tabs simultaneously and read them one at a time. There is now way sarg can know that. But, more to the point, many unrelated pages will download the facebook logo. It serves as a link from the visited page to the facebook page of the site's author. Those links will add up to the total elapsed time.

     
  • RoFz

    RoFz - 2012-08-06
    • status: open --> closed
     
  • RoFz

    RoFz - 2012-08-06

    Hello fmarchal. Thank you for your thorough explanation. It would be of great help.

     

Log in to post a comment.