What can I infer from "show ip route" and similar BGP commands?

Hello NANOG,

I’m a researcher and I was trying to understand the data I collected from
some BGP Looking Glasses. Basically, I was hoping to see if BGP records can
tell me where my university’s provider (AS3701) is peering with its
providers. I issued two BGP queries to Level3’s LGs, one in Seattle and one
in Amsterdam for my school’s prefix. My strong guess was that our provider
(AS3701) peers with Level3 in Seattle. I was hoping to conclude something
like this: if the peering occurs in Seattle, the Seattle LG should reveal
it, but Amsterdam should not.

AS3701 is Nero (Network for Education and Research in Oregon) which I
assume is a small regional AS. I don't think Nero peers with Level3 in
Amsterdam, however, I get this AS for my next hop even when I issue the
command from Amsterdam. On the other hand “car1.Sacramento1” suggests that
the peering happens in Sacramento.

This result makes me think what I get is from a combination of iBGP and
eBGP, which is also apparent from “Internal/External” keywords in the data.
My main issue is that the keywords are not always available. In some other
LG I just get a next hop IP and an AS path. How can I make sure that the
peering information comes from an eBGP peering? I think the next hop IP
might be the answer, right?

I included the results of the command for both LGs here, hopefully somebody
could explain to me

Hi there,

Perhaps this would be easier and help you out:




Thanks Joel for your detailed explanation. It was very informative. I have
been using routeviews for sometime, but given that I could get this amount
of information from other sources, I decided to give this a try.

On another note, do you think there is any value in checking the next hop
IP? I have been checking and it looks as if when the IP is in the AS at the
head of the AS path, the entry is associated with an iBGP record, right? I
just used the ripe stat to map IPs to AS and it always holds when there is
an AS for the next hop IP.

Thanks again for your input.