[brite-users] What does the Internet look like?

Robert P Ricci ricci at cs.utah.edu
Thu Feb 16 17:33:00 EST 2006


Well, I don't want to speak for someone else, but...

A while back, I wanted an answer to the same question. I talked to John
Byers, one of the authors of the BRITE papers. Basically, his answer is
the one you're hoping not to get. :) Basically, there is no right
setting.

I ended up making router-level topologies with the GLP model (I believe
it's the most recent of the ones implemented in BRITE), with the default
parameters.

If you want topologies that are probably as real as you're going to get,
you could try the topologies gathered with RocketFuel:
http://www.cs.washington.edu/research/networking/rocketfuel/
They are topologies gathered from the real Internet. Since the paper was
published, they've been criticized quite a bit (things like MPLS hurt
their accuracy), but they're probably more representative than what
you'll get from a topology generator. However, the information that
comes with them is very incomplete - we found several ISPs to have
disconnected networks, and they don't have latency or bandwidth
information. Still, if you want to go this route, let me know and I'll
see if anyone around here has a rocketfuel -> NS converter.

Thus spake Massimo Canonico on Wed, Feb 15, 2006 at 10:22:45PM +0100:
> Hi,
> I apologize if this question is "off-topic" but in some way is a common 
> problem for the brite's users :
> 
> "What does the Internet look like?"
> 
> I mean, if I would like to set the Brite's parameters in oder to obtain 
> 100 nodes that we could find in Internet..where should I start?
> 
> Is there any paper that suggests when is better to use BA model instead 
> of waxman? How  the bandwidth distribution usually is if you want 
> recreate a Internet scenario?
> 
> I know that probably there is not the right answer, but I would like to 
> have some "starting point" for my study.
> 
> Thanks,
> Massimo.
> _______________________________________________
> brite-users mailing list
> brite-users at cs.bu.edu
> http://cs-mailman.bu.edu/mailman/listinfo/brite-users

-- 
/-----------------------------------------------------------
| Robert P Ricci <ricci at cs.utah.edu> | <ricci at flux.utah.edu>
| Research Associate, University of Utah Flux Group
| www.flux.utah.edu | www.emulab.net
\-----------------------------------------------------------


More information about the brite-users mailing list