Date: Mon, 28 Jul 2003 14:58:18 +1000 From: Mark Hannon <markhannon@optusnet.com.au> To: Luigi Rizzo <rizzo@icir.org> Cc: freebsd-ipfw@freebsd.org Subject: Re: using dummynet to simulate modem, dsl, etc Message-ID: <3F24AD6A.8060504@optusnet.com.au> In-Reply-To: <20030727085734.A89225@xorpc.icir.org> References: <3F212BF7.4060602@optusnet.com.au> <20030725103814.A54554@xorpc.icir.org> <3F23ABF3.1020905@optusnet.com.au> <20030727085734.A89225@xorpc.icir.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Luigi Rizzo wrote:
>> # add serilization of 40 octet ACK to upsteam delay
>> $us_serial_delay = (40 * 8 ) / $$this{us_bw};
>>
>>
>...
>
>
>> # add serilization of MTU octet data to downstream delay
>> $ds_serial_delay = ($$this{mtu} * 8 ) / $$this{ds_bw};
>>
>>
>
>why are you adding this extra delay ? dummynet already does that --
>a packet only comes out of a pipe after what you call the "serialization"
>delay (of the packet itself and any other packet queued in front of
>it).
>
Thanks for clarifying that - for some (probably stupid) reason I thought
that dummynet
simply limited the maximum bandwidth used by a pipe but that it was
still clocked at the
line rate of the original interface. That should improve the results a bit.
Regards/Mark
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3F24AD6A.8060504>