Re: Minutes / Notes

[Joe Abley <jabley@isc.org>]

On Friday, Jul 18, 2003, at 04:12 Canada/Eastern, Pekka Savola wrote:

> ISPs' networks don't crash every other day (if they do, change the 
> ISPs;
> we just can't design protocols to work around broken operational
> practices: that's going to fail, no matter what).  Might happen a 
> couple
> of times during the year, at most, but I'm not convinced that's
> necessarily a huge problem.
This isn't about ISPs' networks "crashing" -- it's about transient 
changes in routing topology, which *do* happen with monotonous 
regularity (circuits fail, or take errors and have to be shut down; 
upstream providers fat-finger routing policy and stopping routes from  
propagating, etc). Also remember that the effects of a re-homing event 
can be felt several levels down; if your path to the DFZ is through 
three intermediate ASes, then you're three-times as likely to feel the 
impact of a re-homing event if the multihoming architecture presents an 
impact to be felt. In some countries, *all* the ISPs are two or three 
ASes away from the core.

The fact that the perception today is that "ISP networks don't crash 
every other day" says more about the functionality of current 
multi-homing practices than it does about the stability of the 
underlying infrastructure.


Joe