[3445] in linux-net channel archive
Re: shaper or whatever
daemon@ATHENA.MIT.EDU (Matthias Urlichs)
Mon Jun 24 07:01:12 1996
To: linux-dev-net@uunet.uu.net
From: smurf@noris.de (Matthias Urlichs)
Date: 24 Jun 1996 12:40:47 +0200
cc: linux-net@vger.rutgers.edu
In linux.dev.net, article <96062431553.014301@drinkel.cistron.nl>,
Miquel van Smoorenburg <miquels@cistron.nl> writes:
>=20
> Ah. Are you basing it on the firewall code? I want to start to implem=
ent
> something based on the firewall code. It adds a bandwidth argument to
> ipfwadm and sets up a bandwidth struct per source/destination specifi=
ed.
Hmmm... but what if you have more than one such entry, and all of them =
need
to share a particular bandwidth?
What happens if there is more bandwidth available than all the bandwidt=
h
structs allocate? Does this limit the available rate or can that be fil=
led?
(That suggests you need to be able to specify two bandwidths -- one
"standard", one "maximum", as both behaviors are useful.)
My own idea is to add an outgoing interface to the firewall rules. That
interface might be an alias interface. Each interface would have a rate
limiter / traffic shaper / whatever-you-call-it, and queue packets to t=
he
"real" interface based on that rate. Incoming traffic would be handled
likewise, i.e. the outgoing-interface of the input firewall rule would
specify the interface (with a separate traffic shaper for incoming traf=
fic)
the packet would 'arrive' through.
Additional advantage: This idea can be used to route packets based on t=
heir
source address (for instance). The forwarding firewall would assign an
interface, and any routing table entries with a different interface wou=
ld
be skipped -> we'd have a working multi-provider router.
Due to limited time to actually do something about this idea, however, =
I've
not done much about it yet. Hopefully that'll change after my vacation.
--=20
You are wasting your time.
--=20
Matthias Urlichs \ noris network GmbH / Xlink-POP N=FCrnberg=
=20
Schleiermacherstra=DFe 12 \ Linux+Internet / EMail: urlichs@nor=
is.de
90491 N=FCrnberg (Germany) \ Consulting+Programming+Networking+etc=
'ing
PGP: 1024/4F578875 1B 89 E2 1C 43 EA 80 44 15 D2 29 CF C6 C7 E0 D=
E
Click <A HREF=3D"http://info.noris.de/~smurf/finger">here</A>. =
42
