Hi,
The debate about frame exchange sequences and what to do with Annex G continues. I am so sorry that I could not be in Waikoloa in person as then it might be easier for me to explain my position. I have decided to try to put my thoughts
to paper so that others can hopefully more easily follow my thoughts, and, of course, react accordingly.
Referring to the infamous Figure 10-14
The steps are:
- The AP sets the FES up by setting protection to cover the entire exchange (FES 3 in the diagram)
- STA 1 sends its BA/Ack and then changes from a participant to a listener. BUT when it replies, it sets the protection so as to cover the remaining part of FES 3. It can be visualized that there was an
FES 1, which is now over, but STA1 must set the protection correctly for overall FES 3.
- STA 1 does now take on a listener role. The point being that it is free to do passive or off-channel things. It is argued that therefore it is now not participating in an FES so it must have been a separate
FES (FES 1). To my mind I don’t care. The point is that STA 1 is not now participating in FES 3.
- Basically, I am saying there is no need to argue over one or three FESs, it is simply a question of whether or when a STA is participating or not in an FES, which is pretty obvious, IMHO.
If we want to make the behavior of a non-participating STA clear, then it should be carried out in the Spec, for example where it states, “once the current frame exchange sequence is complete.” Could be re-written as, “once
participation in a current frame exchange sequence is complete.” This is because the rules were written way before anything like Figure 10-14 was envisaged.
I know that others see a separation between the FES and the protection mechanism(s), but to my mind they are joined at the hip. An FES sets protection and relies on it to send and receive all the packets in the FES. Hence, saying that
there is an FES in progress, or that the protection mechanism is set, is effectively one and the same, in practice. I do not see any reason to change the definition of FES or to use Annex G as a place to explain the subtle differences. I can’t see that would
help, in practice.
I am not sure why we need a new Annex G. The arguments at the moment seem to come down on it being a tutorial, i.e., clicking on an FES in the table to take one to the relevant part of the Spec. Finding a specific exchange can be accomplished
by looking at the Spec “List of Figures” or “Contents”. Also, the present proposal would require updating with every new amendment, an argument we used originally for looking at and changing Annex G.
As I have stated several times, the vast majority of frame exchanges are frame exchange sequences. The basic concept is simple, the FES protects itself by controlling the wireless medium. The rules on what a participating STA cannot do
are pretty obvious, and I believe are fine. BUT it can be argued that what a non-participating STA can do is not clear. We could/should investigate making those rules clear. To my mind that should not be difficult and cannot or should not be done in Annex
G.
I hope this helps.
Graham
To unsubscribe from the STDS-802-11-ARC list, click the following link: https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-11-ARC&A=1