[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: Scott made me do it.

On Tue, Aug 18, 2009 at 8:26 PM, Andrew Lewman <andrew@xxxxxxxxxxxxxx> wrote:

 The summary of results:

1) Native polipo is 54.5% faster on average than native privoxy.  This
could be due to polipo's caching, http 1.1 pipelining, and it can serve
bits as fast as they come in from the network.  Privoxy needs to load
the whole page, scan it, and then send it to the client.  Even if
privoxy filtering is disabled, it still works the same way.

2) Polipo caching shines with Tor usage.  Common images are cached, and
served from the memory cache in single-digit millisecond ranges.
Privoxy needs to wait for Tor to wholly deliver the bits.  Caching is
faster, this we know already.  However, from a user perspective, it's
just faster to load pages.


Constructive criticism is welcome.

Hello, Andrew.

One thing to consider is that polipo's cache maxes out at 32,000 entries.  Once you hit that, nothing more is cached.  And there is no built-in pruning mechanism to keep it smaller. I believe that Juliusz said he would address this someday, but afaik it has not been done yet.

There also is the annoying fact that polipo crashes regularly, which Juliusz knows about, but I understand that he has not had enough time in the last year to fix it.  I believe it is non-trivial to fix.

Lately I have been experimenting with DeleGate proxy cache software, and it has passed almost all testing so far.  It does hang for some reason when trying to do a port 443 connection to https://login.icq.com but this is not a hard crash like polipo - just one of its forked threads goes into a loop.