Re: addressing format questions

[Padmaparna Haldar <haldar@ISI.EDU>]

On Sun, 13 Sep 1998, Lloyd Wood wrote:

> A browse through the comments in ns/tcl/lib/ns-address.tcl was
> enlightening. 
> 
> I see that "mcast should be enabled before Simulator is inited (new
> Simulator is called)" - yet a lot of Tcl sample code doesn't appear to
> do this.

so for backward compatibility's sake 1 bit was forced to be reserved for
mcast.

> 
> I was hoping that something of the form: 
> 
> $ns set-address-format hierarchical 1 10
> 
> would provide a workable solution for me with a customisable flat
> address space of useful size, but alas it provided instead: 
> 
> can't read "cluster_num_": no such variable
>     while executing
> "eval $rl send-num-of-clusters $cluster_num_"
>     (procedure "_o4" line 28)
>     (Simulator hier-topo line 28)
> [..]
> 

while developing support for hierarchical rtg, #levels of hierarchy was
assumed (not quite correctly, as you pointed out) to be more than 1, as
the condition that you are trying to simulate was thought to have been
already supported by means of expand-address and expand-portaddress.
So I'll be working on getting single level hierarchy to work.
Meanwhile I'll also look into why/how expandaddr seem to be breaking with
mcast interaction.

Also, you might want to have hierarchical addressing,and
1. not turn on hierarchical routing for your simulation.
2. default: have hier rtg as well (turned on by "Simulator set
EnableHierRt_ 1" in Simulator procedure "set-hieraddress{}"
(ns-address.tcl).

for option1, the EnableHierRt_ flag would ahve to be set to 0.

You can also take a look at hierarch. example scripts
hier-rtg-{10,100}.tcl in ~ns/tcl/ex.

--Padma.

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

------------------------------------------------
Be true to your work, your word, and your friend.
--Thoreau

Padmaparna Haldar
ISI/USC.
310.822.1511 #352