I am not really an expert in 802.15.4 but I have thought about your question.
I suppose you are assuming a mesh topology, where nodes are allowed to transmit and receive information between them, not only with the PAN coordinator.
In this kind of peer-to-peer topologies, the standard does not specifies anything about ACK's, as it says un section 5.4.2.3.
>From my point of view, there would be no problem in confusing the ACK's. In P2P topoplogy, without using a synchronization mechanism, nodes should "listen" constantly to the channel, and send frames using unslotted CSMA-CA.
These means that if A is sending a message to B, and C wants to transmit at the same time, C would perform a channel sensing. When it detects the ongoing transmission (A<-->B) it should backoff and try again later on. The backoff perdid should be long enough to allow A and B to finish their communication (including ACK's and other messages). Then, if C uses the same DSN, there would no problem of confused ACK's, because A and B would have ended their transmissions.
An ACK frame is only "correct" in the context of the communication established between two nodes.
However, this can be a problem when using differents PAN's. In this case a node could confuse ACK frames, but I think this is covered also in the standard.
A ------------- B
|
|
|
|
C --------------- D
PAN1 = A and B
PAN2 = B and C
PAN3 = C and D
In this case there could be some problems. Let's suppose B sends a message to A and waits for the ACK, and at the same time D sends a message to C and also waits for ACK. This can happen because when D performs channel sensing, it does not "hear" the transmission from B.
If the ACK form A is lost, B could still receive an ACK from C, (destined to D) because it is in his coverage area. If DSN field is the same, a confusion can happen.
I am not an expert of 802.15.4 and I think this aspect is taken on account on the standard, but now I can not remember exactly.
Maybe I am not right, so any remarks, comments or further explanations would be welcomed.
I hope this can help you.
Regards,
Arnau Quintana
On 12/15/05, Martin Gercke <[EMAIL PROTECTED]> wrote:
Hi,
I have a small question about acknowledgments in 802.15.4.
According to the standard they don't hold any address information.
Neither source nor destination.
So, in a PAN with many devices all in radio reach of each other it could
happen that 2 messages are send with the same DSN from 2 different motes
(e.g. A->B and C->D),
messages from A-B gets acknowledged but message from C->D is lost. Since
both messages have the same DSN C could confuse ACK send from B->A with
the one he's waiting for, right?
Martin
_______________________________________________
Tinyos-help mailing list
[email protected]
https://mail.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help
_______________________________________________ Tinyos-help mailing list [email protected] https://mail.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help
