Re: tc, netem: Emulating multiple connections, problem with limit on the number of PRIO bands
- From: Markus Rehbach <Markus.Rehbach@xxxxxx>
- Date: Tue, 28 Aug 2007 20:37:35 +0200
Patrick Reinhardt wrote:
Hi,
to emulate a wireless network I have to assign a packet loss depending on
the target IP address. The idea was to use a PRIO qdisc with n bands for n
nodes and use filters to assign packets to the appropriate band depending
on the target address.
Unfortunately, the number of bands is limited to 16 (TC_PRIO_MAX in kernel
and iproute2). Because of the possible occurence of side-effects of
changing TC_PRIO_MAX to let's say 1000 I am looking for an alternative.
Distributing all packets to about 60 ifb-devices doesn't seem to be a
perfect idea either.
You could build a HTB hierachy and use NETEM as the leaf queuing discipline.
Markus
.
- References:
- tc, netem: Emulating multiple connections, problem with limit on the number of PRIO bands
- From: Patrick Reinhardt
- tc, netem: Emulating multiple connections, problem with limit on the number of PRIO bands
- Prev by Date: Re: Model OSI
- Next by Date: Re: Model OSI
- Previous by thread: tc, netem: Emulating multiple connections, problem with limit on the number of PRIO bands
- Next by thread: learning network protocol testing
- Index(es):
Relevant Pages
|
