[5758] in North American Network Operators' Group
Re: BigISP<-->SmallISP peerings
daemon@ATHENA.MIT.EDU (Vadim Antonov)
Sat Oct 26 16:17:35 1996
Date: Sat, 26 Oct 1996 13:01:58 -0700
From: Vadim Antonov <avg@quake.net>
To: asp@partan.com
Cc: nanog@merit.edu, zachary@zachs.place.org
>> Yes, but then from the point of view of large ISP the peering is of zero
>> value. You see, it has to deliver packets to IXP anyway. OTOH, the
>> load on routers, bloated configurations and engineering resources to
>> support the additional peering are quite real.
>Not quite zero - if you have direct Small/Big vs Small/Transit/Big,
>then you bypass any potential problems with Transit.
> --asp@partan.com (Andrew Partan)
Well, but then you get problems with Small being sloppy with what
they announce. Transit is large and supposedly is clueful enough
(and have resources to develop tools) to ensure some sanity of
routing information.
Overall, the balance of problems is still zero; but the risk that your
other business will be harmed is less in case of transit.
--vadim
