[brite-users] Top-down hierarchical topology generation seems not to work for me!!

Anukool Lakhina anukool@cs.bu.edu
Tue Aug 12 12:32:01 2003


Hi Oliver,

I'm not sure what you mean by arb units.   What exactly are you trying to
accomplish?  

Recall that the HS parameter specifies the side of the big square and the
LS parameter specifies the subdivision of that big square into smaller
squares.   Since you're using the top down generation, you'll build an AS
topology first in the plane specified by the HS and LS units of the AS
model.  Then for each AS, you'll generate a router level topology in
another plane, whose size is dictated by the HS and LS of the router
model. 

Anukool





On 12 Aug 2003, Oliver Holland wrote:

> Hi all,
> 
> Sorry if this is a no-brainer question.
> 
> The (top-down) hierarchical topology generation seems not to work for
> me!! Am I missing some configuration parameters?
> 
> Using the configuration parameters I have specified (see attached file),
> you would expect the ASs to be separated by a max ~1400 arb units (HS
> for ASs specified as 1000 arb units), and the routers in each AS to be
> separated by max ~140 arb units (HS for router level specified as 100
> arb units). But the brite output I get seems to be only using the router
> setting of HS (i.e. only separation of max ~140 arb units for nodes, and
> positioning of all nodes in the topology in a 100*100 plane). Am I doing
> something wrong? Note that I am using the Java exe for generation, but
> the results are the same for c++ exe. The top-down 2-level hierarchy is
> definitely specified in the 'topology type' box at the top of the gui.
> 
> I have included the .brite output for reference.
> 
> Thanks,
> 
> Oliver Holland.
> 
> Ps. I have improved the brite ns2 export format, so that (i) it works
> with the current version of tcl used in ns2 (which is probably very
> similar to the versions that have been used for a least the last couple
> of years), and (ii) it specifies node positions and link orientations in
> ns2, hence you can see the node positions/link orientations correctly in
> nam. Once I have reduced node sizes in nam for the ns2 export files (so
> that the new nam outputs don't only produce nodes that are in the right
> positions, but nodes are also not so large that you can't see what is
> going on!) I will submit it to the brite mailing list.
>