[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Aggregating channels...
Getting consumer broadband operators to provide BGP + the overhead of IP
assignment and AS assingment means that implementing BGP for this is
'difficult'. One way I have considered doing this is to use some sort of
tunnel that sends traffic over two different 'sub tunnels' - but this is
only useful between two fixed point, ala VPN.
I know this sounds arrogant, but if you haven't done BGP, then don't start
with this kind of scenario.
Using IPSec / isakmpd + mods to split traffic over two pairs (or more) of
flows, using different IP address endpoints may be feasible, *and* does not
need to co-operation of the ISP. Lots of work I expect.
Peter
----- Original Message -----
From: "Henning Brauer" <lists-openbsdtech@bsws.de>
To: <tech@openbsd.org>
Sent: Tuesday, December 04, 2001 1:36 AM
Subject: Re: Aggregating channels...
> On Mon, Dec 03, 2001 at 05:17:39PM -0700, Ralph Forsythe wrote:
> > On Tue, 4 Dec 2001, Paul de Weerd wrote:
> > > No, we use this for channel coupling from one location to another
> > > within our own WAN, so we do what Ralph Forsythe mentioned - using
> > > data lines to the same "provider" (ourselves). I just wanted to
> > > go on record that this is possible on a *nix system. I've not seen
> > > what you (and OP) suggest anywhere (and doubt if it's possible).
> > I don't think it's possible to span two ISP providers together either,
the
> > basis of routing won't allow for that.
>
> It IS possible. Requires both (ar all, could be more) to speak BGP with
you,
> you'd had full-mesh BGP tables than (that's abou 110k routing table
entries).
> A nearly equal balancing over the lines isn't that easy even then, though,
> and BGP is for sure nothing for home or normal corporate use.
>
> Greetz
>
> Henning
>
> --
> * Henning Brauer, hostmaster@bsws.de, http://www.bsws.de *
> * BS Web Services, Roedingsmarkt 14, 20459 Hamburg, Germany *
> Unix is very simple, but it takes a genius to understand the simplicity.
> (Dennis Ritchie)