Bug 1809 - lr-wpan features needed
lr-wpan features needed
Status: CONFIRMED
Product: ns-3
Classification: Unclassified
Component: lr-wpan
pre-release
PC Linux
: P5 enhancement
Assigned To: Tom Henderson
:
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2013-12-03 09:35 EST by Tom Henderson
Modified: 2013-12-04 04:55 EST (History)
2 users (show)

See Also:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Tom Henderson 2013-12-03 09:35:18 EST
(copied from ns-developers message from Tommaso Pecorella, Oct 17 2013).

The list of changes I'd like to include in lr-wpan before its inclusion in ns-3 are:
1) packet collisions: 2 packets sent by 2 nodes shall not be received correctly by a 3rd node. One of them have to be discarded.
2) ACKs for data packets
3) callbacks (or traces) for relevant data about packet transmission, i.e., how many retransmissions it needed for successful transmission.
4) NetDevice compliance. 6LoWPAN should be able to hook on lr-wpan.

Other features I'd strongly like and I'm willing to work on:
1) 2011 standard compliance. It's mainly a matter of re-designing an API.
2) Beacons and node association (this is the part I have implemented and I'm not happy with)
Comment 1 Tom Henderson 2013-12-03 14:14:11 EST
LrWpanNetDevice::SetAddress() should handle also extended addresses
Comment 2 Sascha Jopen 2013-12-04 04:55:28 EST
(In reply to Tom Henderson from comment #0)

I already implemented feature number 2 (ACKs for data packets). However, this i still have to verify correct operation. This will happen in the next few days/weeks.

While implementing this feature i partially implemented and fixed feature number 1 (Packet collisions). The LrWpanPhy did not correctly implement its state machine. Sending and receiving at the same time lead to crashes. I think that i fixed this. For packet collisions i currently drop every but the first packet being received. The first one is always correctly received in such a case. However, the signal powers of the other packets should be considered for the first packet as well. Because i do not know how to correctly model this, someone else has to implement it.

I will post my modifications as soon as i have tested them.