[c-nsp] Basic understanding of 6PE and 6VPE

Jared Mauch jared at puck.nether.net
Mon Jun 25 16:35:03 EDT 2012


On Jun 25, 2012, at 4:16 PM, Gert Doering wrote:

> On Mon, Jun 25, 2012 at 04:37:55PM +0200, adam vitkovsky wrote:
>> If the MPLS backbone doesn't have the TTL propagation enabled 
> 
> ... then you're already harming your customers by taking away useful
> diagnostic capabilities.

This isn't the 1990's or even the early 2000's anymore where you have to worry about hosts with a default TTL of 32 anymore.  This was a major problem for some of the content providers back in the day when you couldn't reach them with Win95/98 hosts.

This caused problems for some major search engines "in the day" and their upstream turned on TTL hiding with the newfangled tag^WMPLS switching and it made some hosts able to connect again.  Those days are now over and the teething is mostly gone with MPLS.  The forwarding plane and other defects do continue to exist in a fashion that are confirmation of the law of large numbers. 

Remove the ability to diagnose your network by a junior or outside engineer at your own peril.  Gert was willing to terminate service based on that apparently :)  You may face the same challenge if you follow his path without a "very good(tm)" reason.

- Jared


More information about the cisco-nsp mailing list