Re: node-field-size on >=ns 2.1b3, >128 nodes

[Lloyd Wood <eep1lw@surrey.ac.uk>]

On Fri, 24 Jul 1998, Padmaparna Haldar wrote:

> > I'd expect a complaint about node-field-size of *8* bits rather than
> > 7; 
> 
> 1 bits is allocated for mcast (though this seems like a waste this is done
> to take care of, again, backward-compatibility for the existing scripts),
> so that leaves 7 bits for the nodes and 8 for ports, bu default.

(reads tcl/lib/ns-address.tcl)
That bit is allocated to multicast even if 

Simulator set EnableMcast_ 1

is _not_ issued at all in the script? (I'm not using multicast in
these particular scripts, so didn't need to set it.)

That appears to be where my misunderstanding lies; I was requesting
information to determine the number of nodes, and then using that to
decide whether the expanded address space was necessary or not via
the overly smart:

if { $computedtotalnoofnodes > 256} {
   puts "expanded address space set"
   Node expandaddr
}

(This seemed to work for a number of simulations of networks of 132
nodes for ns 2.1b2 and previous.)

So, that 256 needs to be 128, even when EnableMcast_ is not set to 1,
and the comments in tcl/lib/ns-address.tcl supplement section 3.1 of
the ns docs. (which might say 'by default, all simulations are
restricted to 128 nodes for a uniform addressing space' etc.) 

Thanks,

L.

optimising himself into trouble.

<http://www.ee.surrey.ac.uk/Personal/L.Wood/>PGP<L.Wood@surrey.ac.uk>