| Thread Links | Date Links | ||||
|---|---|---|---|---|---|
| Thread Prev | Thread Next | Thread Index | Date Prev | Date Next | Date Index |
|
Hello everyone, Thanks a lot for the questions you asked in a chat during the presentation – please find my answers below. For those who had no chance to ask the question – please use this thread for that. from [V] Mark RISON Samsung to everyone: 11:51 PM q ((why EIFS for legacy STAs?) from [V] Dmitry Akhmetov, Intel to everyone: 11:53 PM we need a way to slow down all STAs channel access including legacy once. DS signal should be understood by all clients. easiest way is to send _preamble_ only signal
that will be interpreted as error at all STAs including legacy which will trigger EIFS. While “other” STAs run EIFS, the STA(s) that have sent DS signal exercise HiP EDCA within short contention window between 0..7 slots from [V] M. Kumail Haider Meta to everyone: 11:55 PM @Laurent: Does switching to this "prioritized mode" also reset the RBO? Otherwise BO may be longer since these are retransmissions? Dmitry: The switching does not reset RBO, only successful transmission reset RBO. Example: STA suffered from 4 retries, current CW=127 and RBO would be some big number.
STA send DS, announce HiP EDCA, draw new BK=uniform (0..7), compete and transmit. If transmission successful, it reset Cw to CwMin. After successful Hip EDCA period STA use regular/unmodified EDCA with reset CW parameters. It may happen that there is more
than 1 STA participating in HiP EDCA (say STA1 and STA2) -> only one can win contention, so STA1 win -> STA2 did not obtained TXOP in current HiP EDCA round -> STA2 _does_not_ reset CW/retry/etc-> it send DS signal at a next opportunity. from [V] Mark RISON Samsung to everyone: 11:55 PM DS == DSSS? from [V] Laurent Cariou, Intel to everyone: 11:56 PM no, defer signal from [V] Tomoko Adachi Toshiba to everyone: 11:58 PM then, DS signal == defer signal signal? Dmitry: Yes, good catch. DS – defer signal , but it is hard to only say “send DS”, but much easier to say “send DS signal”
😊 from [V] Patrice NEZOU, Canon to everyone: 11:57 PM A STA can transmit low latency flows and non low latency flows at the same time. How can you differentiate the data transmission when using this mechanism ? Dmitry: This is possible, we just need to think on rules (this what Alfred asked as well) when and how to use the mechanism appropriately. from [A] Yanchun Li, Huawei to everyone: 12:00 AM q (The scheme degrades the experience of legacy high priority EDCA queue, e.g. AC-VO. Has QoE of legacy STA been taken into consideration?) Dmitry: yes, legacy coexistence /backward compatibility need to be taken into account and rules how to use DS for LL need to be defined from [V] Chunyu Hu Spreadtrum to everyone: 12:00 AM Q (when the STAs exit the EIFS mode?) Dmitry: as of today – as soon EIFS timer expires, STA exit EIFS and start regular contention from [V] Evgeny Khorov IITP RAS to everyone: 12:03 AM Q (the AP already has ways to preoritize RTA, e. g OFDMA: just send a TF and allocate resources to RTA STAs) why do not you consider? Dmitry: relying on AP and going strictly on centralized access is a bit extreme in our perspective, having “distributed” variant of channel access shall not be overlooked
as Laurent pointed out *in a first slides of the presentation. from [V] Alfred Asterjadhi, Qualcomm Inc. to everyone: 12:03 AM Should be there rules governing the use of hip EDCA/sending DS? Dmitry: yes, the rules need to be developed from [V] Liuming Lu OPPO (privately): 12:10 AM Hi Dmitry, thanks for sharing your document. some Qs: can the DS signal be transmitted by multiple STAs at the same time? Dmitry: yes, this is essentially the idea. We assume that there will be more than 1 STA experiencing collisions OR more than 1 STA with LL traffic in a network. . The
DS signal should be “unified” signal, i.e. like CTSes in a response to the MuRTS. In such a case multiple DS signals transmitted at the same time should be seen as “one” signal from [V] Liuming Lu OPPO (privately): 12:11 AM whether the DS signal collide when transmitted by multiple STAs? Dmitry: They should not if transmitted at the same time – the STAs receiving identical DS signals sent a the same time should see them as one DS signal. I also saw Subir Das Peraton Labs; Chunyu Hu Spreadtrum; Yunbo Li Huawei; Vishnu V Ratnam, Samsung Research ; Gaurav Patwardhan in a chat window Please cast your question here. Dmitry To unsubscribe from the STDS-802-11-TGBE list, click the following link: https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-11-TGBE&A=1 |