On Monday 30 September 2002 12:52, SERBAN Rares wrote:
> Hi Stef,
>
> Please can you detail the arguments to use CBQ and
> HTB? Why HTB is better than CBQ? Which are the
> weakness of CBQ implementation which solved by HTB?
For the performance difference, take a look at the htb homepage. Devik did
some tests with both htb2, htb3 and cbq.
Cbq has a weak theory background. It works by calculating the idle time of
the link. But the only input we have to calculate this is the number of
packets the NIC send. So you have to know your NIC bandwidth, the average
packet size, and some other obscure options (allot). If you have different
packet size or a changing NIC bandwidth (pppoe, pptp), you can get strange
results with cbq.
Htb on the other hand is based on tbf. It does not need to know NIC
bandwidth, nor average packet size. It has a total different approach to do
traffic shaping. I refer to the htb homepage for more info about the theory
background of htb.
And one of the mose convincing arguments to me : htb is actively maintained.
If there is a bug or performance problem, it will get fixed.
Stef
--
[EMAIL PROTECTED]
"Using Linux as bandwidth manager"
http://www.docum.org/
#lartc @ irc.oftc.net
_______________________________________________
LARTC mailing list / [EMAIL PROTECTED]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
