North American Network Operators Group

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

Re: dontaing bgp config files [Re: Risk of Internet collapse grows]

  • From: Ratul Mahajan
  • Date: Mon Dec 02 17:46:05 2002


> 	It is not clear if you mean that tools (e.g. BGP) are
> 	primitive, languages to express policy in BGP are
> 	primitive, or application of what we have (BGP + whatever
> 	language you use) is primitive.  Which is it (or which
> 	subset)?

i would argue all of them; they are so tied to each other that its hard
for me to distinguish. bgp does not let you do everything you want, and at
times lets you do things you don't want. moreover, to my knowledge, the
way most people configure it is also primitive.

but our immediate goal is more modest - trying to understand whats going
on and what the impact of it is. the more challenging task of fixing
things will come later, when we know the current world better.

	-- ratul


On Mon, 2 Dec 2002, David Meyer wrote:

> 	Ratul,
> 
> >> understanding of routing (especially inter-domain) in the research
> >> community is really primitive. this precludes us from having realistic
> >> routing models. we recently started working on understanding prevalent
> >> inter-domain routing policies. the ultimate goal is to improve the
> >> efficiency, robustness and expressiveness of routing protocols.
> >> http://www.cs.washington.edu/research/networking/policy-inference/
> 
> 	It is not clear if you mean that tools (e.g. BGP) are
> 	primitive, languages to express policy in BGP are
> 	primitive, or application of what we have (BGP + whatever
> 	language you use) is primitive.  Which is it (or which
> 	subset)?
> 
> 	Thanks,
> 
> 	Dave
> 
> 
> 
> 	
>