Simple Wireless and RangeLossModel

[Josh Pelkey, 2010-06-11 19:53:52]

Posting message to get this on ns-3-reviews

[Tom Henderson, 2010-06-14 04:28:42]

To clarify, I think that the simple wireless model could go into src/contrib for
this cycle.  It still requires finishing up (no trace sources or helpers
presently).  

IMO, the new propagation loss model is separate from the simple wireless model. 
This model (where interference range equals transmission range) was requested to
facilitate testing of higher layer protocols.

[Nicola Baldo, 2010-06-21 23:39:40]

I think the contribution is potentially useful, because sometimes users need
such a simple wireless channel and device. I agree with Tom that helpers and an
example program should be provided, and that RangePropagationLossModel should be
considered a separate contribution. 

In order for this code to be merged in the main tree, I think we would need to
discuss some related issue:

1) a simple range-based PHY and channel would be of great interest for use with
wifi, so it would be great to find a way to reuse most of the code for this
purpose;

2) the MAC layer functionality of the SimpleWirelessDevice is very similar to
that of AlohaNoackNetDevice, so again we should find a way to reuse some of the
code (e.g., using the interface defined in node/phy-mac.h)


It might still make sense to put the code in contrib as-is (but at least with
helpers and examples) and see if somebody provides some feedback. Still, the
exact policy for stuff to go into contrib is not very clear to me, so I'll be
glad to leave this decision to somebody else.

http://codereview.appspot.com/1587041/diff/1/4
File src/devices/simple-wireless/simple-wireless-channel.cc (right):

http://codereview.appspot.com/1587041/diff/1/4#newcode71
src/devices/simple-wireless/simple-wireless-channel.cc:71: if (distance >
m_range)
I agree that for a simple wireless model it is ok to do this, but then why do
you need a RangePropagationLossModel?

http://codereview.appspot.com/1587041/diff/1/4#newcode78
src/devices/simple-wireless/simple-wireless-channel.cc:78:
Simulator::ScheduleWithContext (tmp->GetNode ()->GetId (), NanoSeconds (3.3 *
distance),
You should consider using a PropagationDelayModel here

http://codereview.appspot.com/1587041/diff/1/6
File src/devices/simple-wireless/simple-wireless-net-device.cc (right):

http://codereview.appspot.com/1587041/diff/1/6#newcode88
src/devices/simple-wireless/simple-wireless-net-device.cc:88: m_rxCallback
(this, packet, protocol, from);
no collision model? Not even a simple protocol interference model?

[AiB, 2012-02-28 11:45:54]

Yea a simple wireless model is required for people doing research in MAC layer
and above who may not have intricate details of the PHY layer involved.

So I think this is very useful model to begin with.