Re: Supporting large very large number of nodes

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

> Date: Fri, 19 Feb 1999 10:42:18 +0000 (GMT)
> From: Randa <randa@comp.lancs.ac.uk>
> To: L.Wood@surrey.ac.uk
> Subject: Re: Supporting large very large number of nodes
> 
> Hello,
> 	I am using 600 nodes in ns. I added the  
> $ns set-address-format expanded after creating the simulator, 
> After running for small amount of time, I found the error message:
> _o33531: no target for slot 2. Have you encountered something like
> that.

Use Node expandaddr. 

As you'll see from:
http://www-mash.cs.berkeley.edu/dist/archive/ns-users/9901/0084.html

I've learned from experience, and later from padma citing some 
(internal cvs?) note not in any ns source I've ever looked at:
$ns set-address-format {anything}
appears to be effectively broken as far as reliability in my
large multicast simulations is concerned...

There are other reasons why that error can occur, but if you're only
writing simple simulation scripts of topologies and events I don't
believe you should encounter them. 

This is something of a reliability problem for ns, if you ask me; at
the least, I'd expect error messages or warnings to appear to indicate
the unreliability of set-address-format, especially since this
objectified command is the 'obvious' and 'natural' one to use.

Note that the current nsDoc.ps doesn't mention set-address-format
or any of its options at all; I'd steer clear of hierarchical
addressing too...

cheers,

L.


> Thanks you and I am looking forward to
> hearing from you soon.
> 
> Best regards,
> Randa
> randa@comp.lancs.ac.uk

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