North American Network Operators Group

Date Prev | Date Next | Date Index | Thread Index | Author Index | Historical

Re: UUNet->Teleglobe

  • From: Richard A Steenbergen
  • Date: Mon Mar 22 09:28:09 2004

On Mon, Mar 22, 2004 at 12:29:39AM -0800, John Obi wrote:
> 
> Hello,
> 
> Do you see any issue between UUNet and TeleGlobe in
> East coast?

No.

> 10  POS6-0.BR3.DCA6.ALTER.NET (152.63.38.117)  82.8 ms
>  82.6 ms  82.6 ms
> 11  204.255.174.198 (204.255.174.198)  84.0 ms  84.0
> ms  84.0 ms
> 12  if-2-0.core2.Newark.Teleglobe.net (64.86.83.213) 
> 348 ms (ttl=236!)  348 ms (ttl=236!)  349 ms

First off, the actual border between UU and Teleglobe is between hops 10
and 11. Since the IP block for the /30 is owned by UU, Teleglobe probably
(either intentionally or unintentionally) did not exchange the PTR info
w/UU. You can tell this a couple of ways:

a) UU's "BR" router designation (Border Router)
b) The lack of the "ix" name on the first Teleglobe interface you see
c) 204.255.174.197 == POS2-3.BR3.DCA6.ALTER.NET

The rest, without knowing anything about Teleglobe's infrastructure in 
particular, looks like a Cisco MPLS-ism. When decrementing ttl through the 
lsp is turned on, you see the latency for the end of the tunnel on every 
reply. Since the rtt is a) consistant for every hop afterwards, and b) 
ends in a place where that rtt makes sense, it is a dead giveaway:

> 18  if-8-0-0.bb1.HongKong.Teleglobe.net (64.86.80.129)
>  352 ms (ttl=240!)  353 ms (ttl=240!)  352 ms
> (ttl=240!)

Happy traceroute engineering.

-- 
Richard A Steenbergen <[email protected]>       http://www.e-gerbil.net/ras
GPG Key ID: 0xF8B12CBC (7535 7F59 8204 ED1F CC1C 53AF 4C41 5ECA F8B1 2CBC)