Question on UL-OFDMA behavior with TXOP=0 and limited trigger frame usage (802.11ax/be)
99 views
Skip to first unread message
hanny h
May 14, 2025, 9:21:08 AM5/14/25
to ns-3-users
Dear ns3 and 802.11 community,
As part of an investigation into UL-OFDMA behavior in accordance with 802.11ax and -be, we have come across an interesting scenario in which we would be very pleased to receive technical information or references to relevant sections of the standard.
Scenario description:
- Two STAs and one AP, all ax/be-capable
- Access Categories used: AC_BK / AC_BE
- TXOP limit is set to 0
- The AP only sends UL trigger frames very occasionally (typically <10 in several seconds)
Nevertheless, both stations show high uplink activity, which is accompanied by a conspicuously high number of collisions and retransmissions (>5 %)
Our previous understanding is based on the assumption that with a TXOP limit of 0, stations in the uplink are only allowed to transmit after receiving a trigger frame. In the UL-OFDMA context in particular, this is the rule according to the standard specifications, as only the AP is responsible for coordinating the uplink resources.
The ns-3 design documentation for Multiuser Transmissions (34.1.3.2.6) contains the following informative note on this:
“Note that, in order for the two frame exchange sequences to be separated by a SIFS (as shown in Figure 'Frame exchange sequences using UL OFDMA'), it is necessary that the transmitting Access Category has a non-zero TXOP Limit, there is enough remaining time in the TXOP to perform the frame exchange sequence initiated by the Basic Trigger Frame and the Multi-User scheduler (described next) chooses to send a Basic Trigger Frame after a BSRP Trigger Frame.”
In my view, this section emphasizes that with a TXOP limit of zero, no consecutive UL-OFDMA signaling sequence can occur unless a new trigger frame is explicitly sent by the AP. This suggests that when TXOP=0, a clearly structured uplink controlled by the AP is expected.
In view of the observations described, the following questions arise for us:
Is any uplink exclusively allowed by trigger frames at TXOP=0, or are there situations where stations are allowed to transmit without explicit assignment?
How is the observed behavior compatible with the standard specifications if stations are permanently active although only a few triggers are sent?
Are there known scenarios or mechanisms (e.g. from practice or simulation) in which the AP deliberately sends trigger frames only rarely?
We would be very grateful for any references to relevant sections from 802.11ax/be, experience from simulation or implementation.
Best regards,
Hannes
As part of an investigation into UL-OFDMA behavior in accordance with 802.11ax and -be, we have come across an interesting scenario in which we would be very pleased to receive technical information or references to relevant sections of the standard.
Scenario description:
- Two STAs and one AP, all ax/be-capable
- Access Categories used: AC_BK / AC_BE
- TXOP limit is set to 0
- The AP only sends UL trigger frames very occasionally (typically <10 in several seconds)
Nevertheless, both stations show high uplink activity, which is accompanied by a conspicuously high number of collisions and retransmissions (>5 %)
Our previous understanding is based on the assumption that with a TXOP limit of 0, stations in the uplink are only allowed to transmit after receiving a trigger frame. In the UL-OFDMA context in particular, this is the rule according to the standard specifications, as only the AP is responsible for coordinating the uplink resources.
The ns-3 design documentation for Multiuser Transmissions (34.1.3.2.6) contains the following informative note on this:
“Note that, in order for the two frame exchange sequences to be separated by a SIFS (as shown in Figure 'Frame exchange sequences using UL OFDMA'), it is necessary that the transmitting Access Category has a non-zero TXOP Limit, there is enough remaining time in the TXOP to perform the frame exchange sequence initiated by the Basic Trigger Frame and the Multi-User scheduler (described next) chooses to send a Basic Trigger Frame after a BSRP Trigger Frame.”
In my view, this section emphasizes that with a TXOP limit of zero, no consecutive UL-OFDMA signaling sequence can occur unless a new trigger frame is explicitly sent by the AP. This suggests that when TXOP=0, a clearly structured uplink controlled by the AP is expected.
In view of the observations described, the following questions arise for us:
Is any uplink exclusively allowed by trigger frames at TXOP=0, or are there situations where stations are allowed to transmit without explicit assignment?
How is the observed behavior compatible with the standard specifications if stations are permanently active although only a few triggers are sent?
Are there known scenarios or mechanisms (e.g. from practice or simulation) in which the AP deliberately sends trigger frames only rarely?
We would be very grateful for any references to relevant sections from 802.11ax/be, experience from simulation or implementation.
Best regards,
Hannes
Reply all
Reply to author
Forward
0 new messages