Re: Importance of HTTP connection keep-alive

[Juliusz Chroboczek <Juliusz.Chroboczek@xxxxxxxxxxxxxx>]

Michael Gersten:

> getting keep-alive to work will help a lot with web browsing,

Fabian Keil:

> Is this an assumption or did you just forget to show your benchmarks
> to back this claim up?

I've just tested this by running

   wget -p http://www.kde.org/screenshots/

That's 87,607 bytes in 14 files -- a small page with a few images.
I'm using tor 0.1.2.8, Polipo 1.0.0, and, for the second series,
Polipo with the attached patch applied.

Polipo was run with no on-disk cache, and was restarted between every
two tests.  The test was run 5 times, and I alternated between the two
versions of Polipo, so there should be no correlation between a given
version and a particular tor circuit.

Average after removing outliers (smallest and largest value):

  Persistent: 33s
  Non-persistent: 61s

Average of all values:

  Persistent: 66s
  Non-persistent: 98s

All times (sorted):

  Persistent: 19 39 40 52 114
  Non-persistent: 44 61 91 92 107

Please feel free to repeat my tests and report the results on this list.

                                        Juliusz

diff -rN -u old-polipo/server.c new-polipo/server.c
--- old-polipo/server.c 2007-04-18 01:01:39.000000000 +0200
+++ new-polipo/server.c 2007-04-18 01:01:39.000000000 +0200
@@ -1661,6 +1661,7 @@
         }
     }
 
+    request->flags &= ~REQUEST_PERSISTENT;
     n = snnprintf(connection->reqbuf, n, bufsize,
                   "\r\nConnection: %s\r\n\r\n",
                   (request->flags & REQUEST_PERSISTENT) ?