Re: Huan Pham: [ns] ns scalability

[Xuan Chen <xuanc@ISI.EDU>]

Hi,

Sorry for the late reply (the original draft of this email was erased
by myself accidently...:().

I have run some simulation which generated quite a lot packets. I hope it
can help you to estimate the running time of yours.

I used a PIII 500MHz PC with 768M RAM for my simulation. The topology is a
simple dumbbell with 3 hops (the sames as that described in
~ns/tcl/ex/dumbbell.tcl, the queueing mechanism is droptail).

In my simulations, there are totoally 21,273,094 end-to-end packets
transmitted in network, which generated about 65M packet-hops. The
simulation runs for about 10 hours.

Compared to the simulation you proposed, I guess it should take arround
800 hours if you use similar PC as I did. But, if you use a 700MHz
machine, the time could probably be reduced to about 500 hours, which is
about 20 days (a long time though...:)).

Hope it helps,
-chen xuan