Should the Scheduling Request (SR) period be a multiple of the TDD pattern period in 5G?

[arman-maghsoudnia]

Dear all,

In the srsRAN implementation, the sr period must be a multiple of the TDD pattern period. My question is whether the spec enforces this limitation or is it just an srsRAN limitation. I thoroughly checked the spec to find the answer to this question, but could not find the answer. I really would appreciate it if anybody can help me with this question.

Best,
Arman

[herlesupreeth]

If I recall correctly, its a srsRAN introduced constraint

[arman-maghsoudnia]

No problem at all. But I was wondering how does the spec handle this challenge. Because after hours of reading the spec, I guess in the spec they do not put this restriction. Do you have any idea about this?

[herlesupreeth]

I think they have put that restriction in a very vague statement (below) in TS 38.213, clause 9.2.4. And, in section 11.1 it deals with TDD pattern etc.

SR transmission occasions in a PUCCH are subject to the limitations for UE transmissions described in Subclause 11.1
and Subclause 11.1.1.

[arman-maghsoudnia]

This is the exact path that I also went down into it to see whether they mentioned what to do or not. But I failed to find what I was looking for. So if anybody could get the answer to the question, I would appreciate it.

[herlesupreeth]

Technically, spec doesn't restrict what SR period to be use. But, I found that if SR scheduled over DL slot, then PUCCH wont be sent.

A UE considers symbols in a slot indicated as downlink by tdd-UL-DL-ConfigurationCommon, or tdd-UL-DL-
ConfigurationDedicated to be available for receptions and considers symbols in a slot indicated as uplink by tdd-UL-
DL-ConfigurationCommon, or by tdd-UL-DL-ConfigurationDedicated to be available for transmissions.

and

For a set of symbols of a slot that are indicated to a UE as downlink by tdd-UL-DL-ConfigurationCommon, or tdd-UL-
DL-ConfigurationDedicated, the UE does not transmit PUSCH, PUCCH, PRACH, or SRS in the set of symbols of the
slot.

[arman-maghsoudnia]

I see, great! Thanks for the help. My brain was so biased looking for the restriction on SR period that I missed this part.
So we can conclude that there are no restrictions on the SR period (if it is in the allowed list for the subcarrier spacing).

periodicityAndOffset : SR periodicity and offset in number of symbols or slots (see TS 38.213 [13], clause 9.2.4) The following periodicities may be configured depending on the chosen subcarrier spacing:

    SCS = 15 kHz: 2sym, 7sym, 1sl, 2sl, 4sl, 5sl, 8sl, 10sl, 16sl, 20sl, 40sl, 80sl

    SCS = 30 kHz: 2sym, 7sym, 1sl, 2sl, 4sl, 8sl, 10sl, 16sl, 20sl, 40sl, 80sl, 160sl

    SCS = 60 kHz: 2sym, 7sym/6sym, 1sl, 2sl, 4sl, 8sl, 16sl, 20sl, 40sl, 80sl, 160sl, 320sl

    SCS = 120 kHz: 2sym, 7sym, 1sl, 2sl, 4sl, 8sl, 16sl, 40sl, 80sl, 160sl, 320sl, 640sl

sym6or7 corresponds to 6 symbols if extended cyclic prefix and a SCS of 60 kHz are configured, otherwise it corresponds to 7 symbols.

For periodicities 2sym, 7sym and sl1 the UE assumes an offset of 0 slots.

But in the case that due to the periodicity, one occasion that requires uplink transmission happens in the DL slot, the UE has to cancel that occasion.

Thanks for the help @herlesupreeth

[frankist]

@arman-maghsoudnia, do you have any particular use-case for a SR period different than the TDD period?

[herlesupreeth]

But I feel it can only be beneficial when the number of UL slots is higher than the DL slots. And I think that you do not support more UL than DL either.

Its supported. You can try TDD pattern DDMUUUUUUU. Also, set

  pusch:
    min_k2: 2
  pucch:
    min_k1: 2
  tdd_ul_dl_cfg:
    dl_ul_tx_period: 10
    nof_dl_slots: 2
    nof_dl_symbols: 10
    nof_ul_slots: 7
    nof_ul_symbols: 0

[arman-maghsoudnia]

I was referring to this discussion:
#252 (reply in thread)
So in the new update the DCI 2_0 and as a result more UL than DL is supported?

[herlesupreeth]

So in the new update the DCI 2_0 and as a result more UL than DL is supported?

DCI 2_0 is not supported but you can change the TDD pattern as I mentioned above which gives you more UL slots than DL

[arman-maghsoudnia]

But if they cannot be scheduled, what would be the benefit of doing so? Please refer to this answer: #252 (reply in thread)

[herlesupreeth]

Just increasing the nof. UL slots will not do any good since DL slots are needed (sending PDCCH for UL DATA) to schedule UL data.

For example: Imagine 1 UE testing scenario and nof. DL slots is 3 and UL slots is 6. In this case the UE can be scheduled to transmit UL data in only 3 slots since the PDCCH to schedule UL data can be sent only over 3 slots so the remaining 3 slots are redundant.

So in my opinion there should be equal number of DL and UL slots.

Sorry, my understanding was wrong back then. It is possible to send multiple UL PDCCHs (over different CCEs for same UE) for scheduling PUSCHs over multiple slots. So, even if there is one DL slot we can schedule all UL slots in a TDD period. Hence, that part has been fixed in the latest release. So, whenever nof. UL slots > nof. DL slots in the TDD pattern configured, we are now using more than one k2 value to schedule all UL slots.