Re: NSClick - performance upgrade for NS-2

Hi all,

I agree that it would be ideal if this functionality could be easily
kept within Click, since this would result in a simulator-independent
solution. However, the thing I liked about the original NS-2 hack of
Michael is it's simplicity. You are sure that you tackle the core
problem: for a given node and a given point in time, it becomes
impossible to have multiple redundant events scheduled. Which was a
real big issue in my specific simulation. And because of the clear and
limited changes in functionality, chances of introducing unwanted
side-effects are quite small.

Personally, i do not have a distinct opinion regarding the specific
way to overcome this scheduling problem in mainline Click. However, i
am convinced that this problem should be tackled, since it quite
strongly limits the practical usability of NSClick when larger
simulations are needed.

Regards,

Wim

Björn Lichtblau wrote:

Hello,

On 19.01.2012 21:37, Michael Neufeld wrote:

Hello-
It's been a great while since I did much with the ns interface to
click, but here's my 0.02. I'd also lean towards doing the work on the
Click end of things - if I'd done it that way in the first place (and
integrated the change) there'd be no need to duplicate the work in
both the ns-2 and ns-3 specific code.

Sure, if it's easy do it there. But if complicated in click and simple
for the sim i'd prefer that, a global map is maybe too much.
Just thinking in the other direction, scheduling and bookkeeping of
events at top-level is already part of a simulator, don' t do that again
in click when it is run within. Duplicating the information which has
been passed to the sim and hoping that both sides stay in sync is what
i'd like to avoid.

The sim interface could probably do with a bit of sprucing up (in
addition to documentation) at this point - quite likely some of the
assumptions I made when designing/coding it aren't true anymore. I'm
not swimming in free time for Click work right now, but can at least
get started on documenting it as it stands and think a bit about what
could be improved.

The sim coupling code is great regarding that it is still easy to use
even for other simulators, i like it!

2012/1/19 Eddie Kohler<[email protected]> :

But I think we could do that without an NS patch, by changing Click
(lib/routerthread.cc), so that it doesn't call
"simclick_sim_send(...SCHEDULE...)" unless it knows there is no timer
scheduled.

I slightly prefer to change the Click side, but only slightly. The
division of functionality between Click and the NS driver is pretty ad hoc.

I don't see how this helps when there suddenly is an earlier event in
click to pass to the sim, although there is already another timer
scheduled. And i can't see an easy way around this for click. Either
extend the simclick interface so click can ask what is scheduled or have
a bookkeeping of what has already passed to the sim somehow. Not?

Regards, Björn

_______________________________________________
click mailing list
[email protected]
https://amsterdam.lcs.mit.edu/mailman/listinfo/click

--
Wim Vandenberghe
Department of Information Technology
Internet Based Communication Networks and Services (IBCN)
Ghent University - IBBT
Gaston Crommenlaan 8 (Bus 201), B-9050 Gent, Belgium
T: +32 9 33 14946 ; T Secr: +32 9 33 14900
F: +32 9 33 14899
E: [email protected]
W : www.ibcn.intec.UGent.be

_______________________________________________
click mailing list
[email protected]
https://amsterdam.lcs.mit.edu/mailman/listinfo/click