--- Nick Tarleton <[EMAIL PROTECTED]> schrieb:
> On Monday 11 August 2003 08:05 pm, Toad wrote:
> > It would be so inaccurate as to be of little use
> for simulating
> > NGRouting, which is deeply sensitive to network
> conditions. OTOH it
> > could maybe yield some interesting results
> elsewhere.
> Then simulate network conditions! :-P
> 
My feelings exactly.  I've been looking around for a
tool to do this and found NIST Net, which is a Linux
kernel patch that lets you simulate things like
loss,latency, and bandwidth on the IP level.
http://dns.antd.nist.gov/itg/nistnet/

Unfortunately by the looks of it, it doesn't let you
simulate the typical up/down bandwidth bottlenecks
we're dealing with.

I'd be willing to write a deamon that would listen to
1000 ports and forward those to other <IP,ports> and
apply simple delay and bandwidth limits at the TCP
level.  

If someone has a LAN with a couple of machines you
could set all the nodes up as if they were behind a
firewall/NAT with address X.  We could then put this
little monster on a macine with address X and have it
forward to all the nodes appling limits.  Does this
sound like it would work?

This would still leave the other end of simulation
open.  We have to set something up which requests and
inserts like real people do.  Do we have to simulate
Frost usage?

__________________________________________________________________

Gesendet von Yahoo! Mail - http://mail.yahoo.de
Logos und Klingelt�ne f�rs Handy bei http://sms.yahoo.de
_______________________________________________
devl mailing list
[EMAIL PROTECTED]
http://hawk.freenetproject.org:8080/cgi-bin/mailman/listinfo/devl

Reply via email to