[squid-users] tproxy performance overhead ?

Hello All,

Does anyone have any numbers on how much overhead tproxy will add ?
Basically, how much drop in performance would we see if tproxy is
enabled.

I imagine that tproxy will have to do some lookup and hence should
incur some overhead.
I also looked at the tproxy mailing list
https://lists.balabit.hu/pipermail/tproxy/2005-February/000167.html
and it seems to suggest that tproxy is not doing very well.

I am also trying a setup myself.

(192.168.1.201)
    client ----------- proxy ----------- server (10.51.6.197)
                eth0 (10.51.6.102)
               eth1 (192.168.1.102)

* There is no direct link between the server and the client.
* The servers' default gw is the proxy, and the proxy has forwading on.
* Polymix4 running on the client & server.
* I dont have any sticky load balancer or routers, hence i am doing this setup.

Do you think this setup will be able to test the tproxy fairly ?

I would appreciate any comments.

Thanks
-- Pranav

------------------------------
http://pd.dnsalias.org
Received on Wed Dec 06 2006 - 14:39:45 MST