Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

[STDS-802-11] TGme topic for WG: aRxPHYStartDelay



--- This message came from the IEEE 802.11 Working Group Reflector ---

Hi all

 

CIDs 1059 and 2109 observe that aRxPHYStartDelay is defined differently in different PHY clauses and even differently for different PHY formats defined in the same PHY clause. Meanwhile aRxPHYStartDelay is used in the MAC clauses as if it were a single constant.

 

A)

One option is for the PHY to expose all values of aRxPHYStartDelay to the MAC and for the MAC to select the maximum aRxPHYStartDelay among the values relevant to the current context. This approach is defined in 22/576 under CID 1059.

 

This is a simple fix but (as Youhan points out) suffers from the problem that aRxPHYStartDelay can be very high for HE/EHT MU PPDUs (hundreds of usec, which is longer than the duration of two short PPDUs) due to the HESIGB/EHTSIG fields, yet oftentimes the MAC clauses uses aRxPHYStartDelay to determine if the PHY is actively processing a PPDU or not, in order to determine if the MAC can transmit something else. Thus a short value of aRxPHYStartDelay is generally preferred.

 

B)

Another set of options attempts to provide a smaller value (or a smaller set of values) of aRxPHYStartDelay:

  • For PHY formats with more than one SIG field (or the existence of subsequent SIG field(s) is indeterminate for a period of time), the PHY issues two primitives:
    • a new first primitive that indicates one of (TBD):
      • The PHY has successfully decoded the first SIG field and so is reporting that a PPDU is present
      • The PHY has successfully decoded a SIG field and so will keep CCA as busy for the duration indicated by the SIG field (or similar)
    • a final primitive that reports the complete RXVECTOR after receiving all SIG fields (i.e., the traditional PHY-RXSTART.indication)
  • The two primitives could be named using one of the following schemes (TBD):
    • Both are called PHY-RXSTART, with a new primitive-index parameter (along the lines of “1” and “2”; or “first” and “last”)
    • PHY-RXSTART-FIRST-SIG.indication and PHY-RXSTART.indication
    • … etc
  • Given we still have DSSS/CCK, HT-GF and LSIG-style PHY formats, then it seems likely that, in some bands, we still have multiple aRxPHYStartDelay’s so option A is probably still a part of the solution.

 

Any comments / concerns / suggestions / recommended directions are welcome.

 

Best regards

Brian


To unsubscribe from the STDS-802-11 list, click the following link: https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-11&A=1