Osmocom Developer Conference 2024

[brendan-mcauliffe]

@herlesupreeth attended OsmoDevCon2024 in May, where he presented "srsRAN Project - Application in UL heavy traffic scenarios".

The Osmocom Developer Conference is an annual event where Osmocom developers from various Osmocom projects and invited speakers meet. Supreeth's presentation goes through some of the most recent developments in srsRAN Project and demonstrates how to configure and use srsRAN Project for high uplink traffic scenarios such as automotive test tracks, media broadcasting, and AR/XR services.

You can find a recording of his presentation here.

The PDF of his presentation can be downloaded here.

To achieve the configuration shown in the demo, the following parameters can be added to your configuration file:

cell_cfg:
  pdsch:
    olla_target_bler: 0.1
    mcs_table: qam256
  pusch:
    olla_target_bler: 0.1
    mcs_table: qam256
    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

[fllay]

Does this configuration solve this issue #252?

[herlesupreeth]

Yes. It does

[andergalvaan]

5G_Test_Files.zip
Hello @brendan-mcauliffe @fllay @herlesupreeth

I have a 5G network deployed with the following setup:

• Modem: Quectel RM520N-GL 5G
• RAN: srsRAN CU/DU with USRP B210 (Split 8 configuration)
• Core: Open5GS

Deployment details:

• srsRAN is running on a physical server with IP 10.98.1.155
• Open5GS is running on a separate physical server with IP 10.98.1.133
• The 5G modem (Quectel RM520N-GL) is connected to a BeagleBone board

I have attached the following files:

• YAML configuration files for both CU and DU
• iperf3 test results performed over the UPF of the 5G Core (Open5GS)

I would like to know if there is a way to achieve the uplink throughput shown in the video with my current setup, or at least improve my uplink throughput. Would it be more effective to use a sampling rate of 30.72 MS/s, MIMO 2x2, and a channel bandwidth of 30 MHz?

Any advice or suggestions to optimize uplink throughput would be greatly appreciated.

Best regards,
Ander

[herlesupreeth]

@andergalvaan I dont see any issue with your configuration. I would suggest to stick with SISO + 40 Mhz BW. Try to use the latest code from main branch (commit id d5fa4f0). I was able to get 120 mbps in UL on my iPhone 13 pro

--== srsRAN gNB (commit d5fa4f0ecb) ==--

Lower PHY in quad executor mode.
Available radio types: uhd.
[INFO] [UHD] linux; GNU C++ version 11.2.0; Boost_107400; UHD_4.1.0.5-3
[INFO] [LOGGING] Fastpath logging disabled at runtime.
Making USRP object with args 'type=b200'
[INFO] [B200] Detected Device: B210
[INFO] [B200] Operating over USB 3.
[INFO] [B200] Initialize CODEC control...
[INFO] [B200] Initialize Radio control...
[INFO] [B200] Performing register loopback test... 
[INFO] [B200] Register loopback test passed
[INFO] [B200] Performing register loopback test... 
[INFO] [B200] Register loopback test passed
[INFO] [B200] Setting master clock rate selection to 'automatic'.
[INFO] [B200] Asking for clock rate 16.000000 MHz... 
[INFO] [B200] Actually got clock rate 16.000000 MHz.
[INFO] [MULTI_USRP] Setting master clock rate selection to 'manual'.
[INFO] [B200] Asking for clock rate 46.080000 MHz... 
[INFO] [B200] Actually got clock rate 46.080000 MHz.
Cell pci=1, bw=40 MHz, 1T1R, dl_arfcn=650000 (n78), dl_freq=3750 MHz, dl_ssb_arfcn=649056, ul_freq=3750 MHz

N2: Connection to AMF on 172.22.0.10:38412 completed
==== gNB started ===
Type <h> to view help
t

          |--------------------DL---------------------|-------------------------UL----------------------------------
 pci rnti | cqi  ri  mcs  brate   ok  nok  (%)  dl_bs | pusch  rsrp  ri  mcs  brate   ok  nok  (%)    bsr     ta  phr
   1 4601 |  15 1.0    2   3.7k    9    0   0%      3 |  24.1 -17.4   1   24    30k    5    0   0%      0     3n   20
   1 4601 |  15 1.0    8    20k   29    0   0%      0 |  25.4 -16.5   1   28    26k    6    0   0%      0    51n   20
   1 4601 |  15 1.0   27    87k   24    0   0%      0 |  26.1 -15.7   1   27   121k   20    4  16%      0   258n   21
   1 4601 |  15 1.0   26   4.8k    3    0   0%      0 |  26.2 -15.7   1   27   9.7k    2    1  33%      0   265n   21
   1 4601 |  15 1.0   24   8.4k    5    0   0%      0 |   n/a   n/a   1    0      0    0    0   0%      0   262n   21
   1 4601 |  15 1.0    0      0    0    0   0%      0 |   n/a   n/a   1    0      0    0    0   0%      0   250n   21
   1 4601 |  15 1.0   26   240k   31    0   0%    119 |  27.7 -13.5   1   27    75k   14    0   0%      0   273n   21
   1 4601 |  15 1.0   26    14k    8    0   0%      0 |  27.8 -13.4   1   27   9.7k    2    0   0%      0   248n   20
   1 4601 |  15 1.0   26   7.3k    4    0   0%      0 |  28.4 -12.9   1   27    19k    4    0   0%      0   279n   20
   1 4601 |  15 1.0    0      0    0    0   0%      0 |   n/a   n/a   1    0      0    0    0   0%      0   269n   20
   1 4601 |  15 1.0   27    39k   21    0   0%      0 |  26.3 -13.8   1   27   8.8M  108    3   2%   700k   276n    7

          |--------------------DL---------------------|-------------------------UL----------------------------------
 pci rnti | cqi  ri  mcs  brate   ok  nok  (%)  dl_bs | pusch  rsrp  ri  mcs  brate   ok  nok  (%)    bsr     ta  phr
   1 4601 |  15 1.0   26   363k  200    0   0%      0 |  26.7 -13.4   1   27   116M 1345   55   3%   700k   278n    7
   1 4601 |  15 1.0   26   363k  200    0   0%      0 |  26.7 -13.5   1   27   118M 1367   33   2%   700k   279n    6
   1 4601 |  15 1.0   26   363k  200    0   0%      0 |  26.8 -13.3   1   27   119M 1379   21   1%   700k   283n    6
   1 4601 |  15 1.0   26   363k  200    0   0%      0 |  26.9 -13.3   1   27   120M 1381   19   1%   700k   285n    6
   1 4601 |  15 1.0   26   363k  200    0   0%      0 |  26.9 -13.3   1   27   119M 1374   26   1%   700k   288n    6
   1 4601 |  15 1.0   26   363k  200    0   0%      0 |  26.9 -13.3   1   27   119M 1375   25   1%   700k   282n    6
   1 4601 |  15 1.0   26   363k  200    0   0%      0 |  26.9 -13.3   1   27   120M 1385   15   1%   700k   258n    6
   1 4601 |  15 1.0   26   363k  200    0   0%      0 |  26.9 -13.3   1   27   120M 1381   19   1%   700k   271n    6
   1 4601 |  15 1.0   26   363k  200    0   0%      0 |  26.7 -13.6   1   27   117M 1348   52   3%   700k   266n    6
   1 4601 |  15 1.0   26   363k  200    0   0%      0 |  26.8 -13.5   1   27   115M 1328   72   5%   700k   269n    6
   1 4601 |  15 1.0   26   363k  200    0   0%      0 |  26.9 -13.3   1   27   114M 1318   82   5%   700k   272n    6

[herlesupreeth]

Try to use the below gain values

  tx_gain: 74                                                   # Transmit gain of the RF might need to adjusted to the given situation.
  rx_gain: 36                                                   # Receive gain of the RF might need to adjusted to the given situation.

With these values I was able to achieve 121 mbps with Quectel RM520N-GL modem

          |--------------------DL---------------------|-------------------------UL----------------------------------
 pci rnti | cqi  ri  mcs  brate   ok  nok  (%)  dl_bs | pusch  rsrp  ri  mcs  brate   ok  nok  (%)    bsr     ta  phr
   1 4602 |   7 1.0   11   136k  200    0   0%      0 |  30.3 -16.5   1   27   121M 1400    0   0%   700k   233n   -7
   1 4602 |   6 1.0   10   127k  200    0   0%      0 |  30.3 -16.4   1   27   121M 1400    0   0%   700k   233n   -8
   1 4602 |   6 1.0    9   125k  200    0   0%      0 |  30.2 -16.6   1   27   121M 1400    0   0%   700k   233n   -8
   1 4602 |   6 1.0   10   130k  200    0   0%      0 |  30.1 -16.8   1   27   121M 1400    0   0%   700k   233n   -8
   1 4602 |   6 1.0   11   136k  200    0   0%      0 |  30.3 -16.6   1   27   121M 1400    0   0%   700k   233n   -8
   1 4602 |   7 1.0   11   147k  200    0   0%      0 |  30.2 -16.8   1   27   121M 1400    0   0%   700k   222n   -9
   1 4602 |   7 1.0   11   145k  200    0   0%      0 |  30.1 -16.9   1   27   121M 1400    0   0%   700k   202n   -9
   1 4602 |   6 1.0   11   137k  200    0   0%      0 |  29.1 -17.9   1   27   121M 1400    0   0%   700k   204n  -12
   1 4602 |   6 1.0   11   139k  200    0   0%      0 |  28.9 -18.9   1   27   121M 1400    0   0%   700k   204n  -10
   1 4602 |   6 1.0   11   138k  200    0   0%      0 |  29.0 -18.8   1   27   121M 1400    0   0%   700k   206n  -11

[andergalvaan]

attached_files.zip

Hello @herlesupreeth ,

I updated to the commit ID you mentioned, and some issues have appeared.

• CU: The error message INI was not able to parse cu_up.nru.++ appeared.
  To resolve this, I modified the configuration file by replacing cu_up.nru.bind_addr with cu_up.f1u_socket.bind_addr. I am attaching the updated configuration file and a screenshot showing the CU working after this change.
• DU: Similarly, the error INI was not able to parse nru.++ appeared.
  I resolved it by changing nru.bind_addr to f1u.socket.bind_addr in the DU configuration file. I also had to update metrics.enable_json_metrics to metrics_enable_json. I am attaching the updated configuration and a screenshot of the DU working after these changes.

Once both the CU and DU were up and running and successfully connected to the Open5GS 5GC, I attempted to connect the Quectel module using the same script that previously worked. However, the device does not connect to the 5G network anymore.

I'm attaching:

• A screenshot of the AMF logs from Open5GS during the connection attempt.
• A packet capture (pcap) taken on the Open5GS side during the Quectel module's connection attempt.
• The script I'm using to connect the Quectel module.

From what I can tell, the issue may lie in the PDUSessionResourceSetupResponse, but I'm not certain.

Since the only changes I made were updating to the new commit and adjusting the configuration as mentioned above, I'm unsure what’s causing the issue now. If you are not sure either, I would really appreciate it if you could guide me on how to revert to the previous commit, so I can test whether modifying rx_gain and tx_gain in the earlier version improves the UL throughput.

Thanks in advance for your help.
Best regards,
Ander

[herlesupreeth]

Looks like in the latest code gNB is rejecting PDU session setup. I am guessing gNB is expecting qos configuration i.e. pass this configuration file https://github.com/srsran/srsRAN_Project/blob/main/configs/qos.yml in addition to existing configuration while running gNB and give it a try. This is quite strange because in the previous versions of code 5QI=9 was supported by default.

If the above does not work then, if you remember the commit on which you were before you could do

git checkout <COMMIT_ID>

and then recompile

OR,

you if dont remember the commits id then you can use the release "release_24_10_1"

git checkout e5d5b44

and then recompile

[andergalvaan]

Hello @herlesupreeth ,

I’ve tried deploying the gNB using the qos.yml file, but unfortunately, the issue persists. Due to this, I decided to revert to commit e5d5b44, and the problems have disappeared. At the moment, after setting the tx_gain and rx_gain to 80 and 36 respectively, I’m achieving approximately 60 Mbps.

Thank you for your support, and please let me know if there are any updates.
Best regards,
Ander

[andrepuschmann]

@andergalvaan Please create a new issue mentioning the commits that work/don't work and share all configs and logs please. Feel free to link this thread here in the issue but the discussion section is not the right place to discuss this.

[andergalvaan]

@andrepuschmann I will open a issue tomorrow with all configs and logs.

[andergalvaan]

I have opened a new issue (#1251). Looking forward to your response! Thanks!

[ujjwalpawar]

You mentioned support for Intel Accelerators. How can I compile for ACC100?

[brendan-mcauliffe]

Hi @ujjwalpawar, hardware acceleration support is only available in srsRAN Enterprise.