They are prepending routes.
Looks like both 43022 are prepending, As well as 47359...Multiple times... They do this to make that route look "bad" so it comes in other transit they have.
Nick Olsen
Network Operations
(855) FLSPEED x106
thanks all, this makes sense now. and i just showed the internet how ignorant I am…
i have my maxas-limit set to 10 based on an article I was reading. perhaps I should up that a bit.
what sort of problems are associated to overly long AS paths? is it more of a system resource control setting?
-g
From: Greg Whynott [mailto:Greg.Whynott@oicr.on.ca]
Sent: Wednesday, November 10, 2010 3:40 PM
To: nick@brevardwireless.com
Cc: nanog@nanog.org
Subject: Re: AS path question.thanks all, this makes sense now. and i just showed the internet
how ignorant I am.
No such thing! We are all here to learn, and so there are no such things as
stupid questions IMO. I certainly won't rail you for asking a question like
that... ahem, Neils Bakker might so there are no guarantees 
i have my maxas-limit set to 10 based on an article I was reading.
perhaps I should up that a bit.
I think 10 is way too short. I would definitely up that.
what sort of problems are associated to overly long AS paths? is it
more of a system resource control setting?
I remember back in 1999 at UUnet we had a bug in IOS in which some AS Paths
were supposedly being prepended well above 255 AS Paths... IIRC that is
above the maximum field length (I'd have to dig it up to be sure though) and
it was actually causing routers to crash as they were interpreting the BGP
messages as malformed. Today we have better protections in IOS preventing
this and other associated "buffer overflow" issues, but at the time there
were no such protections. The very next day we received a patched IOS from
Cisco which contained the maxas-limit commands.
Stefan Fouant
[... learning about path prepending ...]
[snip]
i have my maxas-limit set to 10 based on an article I was reading.
perhaps I should up that a bit.
That article was deeply mistaken. 50 was reasonable for older IOS with
bugs back in ... 2001-2003? I think. under the auspices of apnic, gih
has done serious study of network diameter [specifically in relation to
churn and background noise, but useful data here]. I'm a big fan of
belts and suspenders but would reccomend taking at least a coarse
analysis of the path lengths seen in your neck of the woods from your
providers before setting a number.
Cheers,
Joe