[499] in North American Network Operators' Group
Re: filtering long prefixes
daemon@ATHENA.MIT.EDU (bmanning@ISI.EDU)
Sun Sep 24 14:47:02 1995
From: bmanning@ISI.EDU
To: michael@junction.net (Michael Dillon)
Date: Sun, 24 Sep 1995 11:43:32 -0700 (PDT)
Cc: smd@icp.net, jbradley@westel.com, smd@sprint.net, ejw@cyberstore.ca,
nanog@MERIT.EDU
In-Reply-To: <Pine.LNX.3.91.950921195534.16162C-100000@okjunc.junction.net> from "Michael Dillon" at Sep 21, 95 07:59:47 pm
Resent-From: nanog@MERIT.EDU
>
> Are you suggesting some sort of exchange point or NAP specifically to
> break out longer prefixes from shorter prefixes that cannot be
> topologically aggregated? Would something like this enable people
> to maintain provider independent addressing (i.e. no renumbering) by
> merely paying a fee to an exchange point that is well connected and
> settling for less optimal routing?
>
> If this will work in practice, it seems like the perfect tradeoff. On the
> one hand you must renumber when changing providers but you get optimal
> routing. On the other hand, you avoid renumbering but you pay a few bucks
> and have less than optimal routing.
>
> Am I missing anything here?
>
Perhaps a couple of things:
- Common transit agreements for all particpants
- Single point of attachment, ie you must renumber
if you home to another provider or exchange.
--
--bill
