R1-2508686
discussion
Discussion on 6GR Channel Coding
From Xiaomi
Summary
Xiaomi presents 2 observations and 4 proposals on 6GR channel coding, questioning the need to optimize latency and error floor given unchanged 5G requirements. They advocate reusing NR LDPC and polar codes within NR range, require rigorous justification for any specification changes beyond NR range, and propose channel-wise rather than scenario-wise evaluation methodologies.
Position
Xiaomi questions the motivation to optimize decoding latency and error floor performance for 6GR, arguing the relevant 6GR requirements (user plane latency of 1ms for hRLLC and 4ms for IC, reliability of 10^-5 for hRLLC) are identical to 5G NR uRLLC and eMBB. They support reusing NR LDPC design for data rate within NR range and NR polar code for control information within NR range (larger than 11 bits). For beyond NR range, Xiaomi requires thorough and vigorous evaluation on memory cost, computational complexity, and hardware complexity before justifying any incremental LDPC enhancement with specification change, and prefers straightforward lifting size extension over proto-matrix/BG structure changes. They propose channel-wise evaluation methodology (data channel LDPC for IC/hRLLC/MC, control channel polar for IC/hRLLC/MC) instead of the 5G scenario-wise approach, and recommend BLER simulation targets down to 10^-3 for IC/MC and 10^-4 for hRLLC including a 10^-5 target for XR service reliability.
Key proposals
- Observation 1 (Sec 2): The motivation to optimize decoding latency and error floor performance is questionable given the relevant 6GR requirements are the same as 5G.
- Observation 2 (Sec 2.1): Implementation means exist to achieve 200Gbps peak data rate, if needed, using the 5G NR LDPC.
- Proposal 1 (Sec 2.1): For within NR range, confirm the working assumption for 6G LDPC and 6G polar — for data rate within NR range, reuse of NR LDPC design is supported; for control information within NR range (larger than 11 bits), reuse of NR Polar code design is supported.
- Proposal 2 (Sec 2.1): For beyond NR range, thorough and vigorous evaluation is needed to justify peak data rate oriented incremental enhancement with specification change for LDPC; designs motivated by increase of UCI/DCI payload size(s), if any, shall be triggered by relevant discussion instead of channel coding discussion.
- Proposal 3 (Sec 2.2): For 6GR, channel coding evaluation shall be performed channel wise instead of scenario wise — LDPC is the data channel candidate and evaluation assumptions need to reflect requirements for at least IC/hRLLC/MC; Polar is the control channel candidate and evaluation assumptions need to reflect requirements for at least IC/hRLLC/MC.
- Proposal 4 (Sec 2.2): For 6GR, BLER simulations down to 10^-1, 10^-3 is recommended (to observe the error floor) for IC/MC; BLER simulations down to 10^-4 is recommended (to observe the error floor) for hRLLC.