[brite-users] which model? Any converter?
anukool at cs.bu.edu
Tue Jun 22 22:11:44 EDT 2004
To answer your first question, selecting which model is the most
representative is really an open research question. You are correct: the
BA and GLP models in BRITE are more geared towards AS level modeling. If
you're interested in using router level topologies corresponding to an AS,
you're probably best off using the rocketfuel maps directly in your
Second, unfortunately, I don't know of any means to convert all the
various topology formats. But if you write such a tool, it would be
really be useful to many in the community, so please do release it! :)
Sorry I can't be more helpful.
On Mon, 21 Jun 2004, Fetahi wrote:
> We have a simulator that takes in network topologies in the form of .brite
> files. And we want to run our simulation on a topology of routers within an
> I have two questions. First, I am unable to determine which model generates
> "the best" router level topology for a network in an AS. I know I have to
> use one of the power law models [BA, AB or GLP] and I also know that AB is
> the an extension of BA and GLP is an improvement of AB. But it looks like
> these models are geared more towards modeling AS level topology instead of
> router level topology. So which model is "the best" for our case?
> Another question is, we are also planning on using real network topologies
> for our simulator in addition to the synthetic topologies. For this, we are
> planning on using the ISP maps generated by Rocketfuel. However the data is
> available in .cch format and obviously incompatible with the .brite format.
> A little search and I found that some of the data in Rocketfuel is available
> in .gml and .ns2 formats. However I am unable to find any converter that is
> capable of changing files in .cch, .gml or .ns2 format to .brite. I even
> tried to locate a converter that is capable of changing form the above
> formats to the GT-ITM and INET formats (to use brite to change them to
> .brite) but still I was unable to find one. Before I try to write a
> converter on my own, does anyone has a way out, or know of any converter
> that is capable of solving the above?
> brite-users mailing list
> brite-users at cs.bu.edu
More information about the brite-users