Unit of DRB.RlcDelayUl

[ersuneetsingh]

I’m using ORAN-SC RIC to monitor the RAN, and I generate UDP traffic using iperf to observe uplink RLC delay (RlcDelayUl) at 1-second intervals. However, I'm noticing that the reported RlcDelayUl values range between 700–1200 ms, which seems unusually high — I was expecting something in the range of 10–50 ms.

I also observed that the unit for DRB.RlcSduDelayDl is 0.1 ms, whereas for DRB.RlcDelayUl it's labeled simply as ms. So I’m unsure: do I need to multiply the RlcDelayUl values by 0.1 as well? Or are the values already in milliseconds?

If the units are correct as-is, I’m not sure what could be causing such high RLC delay. Could anyone provide insights or suggestions?

[ersuneetsingh]

This is a bit urgent—any insights or suggestions at your earliest convenience would be greatly appreciated.

Do the DRB.RlcDelayUl values require scaling by 0.1, or are they already in milliseconds as reported?

[pgawlowicz]

Hi,
The reported unit is in ms. You are right, the values between 700-1200ms seem strange. could you post your gnb logs?

[ersuneetsingh]

gnb.log

Hi,
Thank you for the information. I’ve attached the gnb.log file. I used iperf3 to send UDP packets at a throughput of 1 Mbps and captured an RLC uplink delay of around 8722 ms. Even when I used ping with the default packet size (8 bytes), the RLC delay was around 1 to 2 ms. However, when I set the packet size to 1200 bytes for ping, the RLC delay increased to around 8000 ms, while the ping latency remained around 35 ms (ICMP is not included in log). As per the below RLC value, this shows it should be in us, but as you mentioned the unit is in ms, this is strange.

-Measurements Data:
--UE_id: 1
---granulPeriod: 1000
---Metric: DRB.UEThpDl, Value: [0]
---Metric: DRB.UEThpUl, Value: [1024]
---Metric: RRU.PrbTotDl, Value: [0]
---Metric: RRU.PrbTotUl, Value: [0]
---Metric: RRU.PrbAvailUl, Value: [105]
---Metric: RRU.PrbAvailDl, Value: [106]
---Metric: DRB.PacketSuccessRateUlgNBUu, Value: [100]
---Metric: DRB.RlcSduDelayDl, Value: [100]
---Metric: DRB.RlcSduTransmittedVolumeDL, Value: [0]
---Metric: DRB.RlcSduTransmittedVolumeUL, Value: [1014]
---Metric: DRB.RlcPacketDropRateDl, Value: [0]
---Metric: DRB.RlcDelayUl, Value: [8696.0]
---Metric: DRB.AirIfDelayUl, Value: [45.0]
---Metric: CQI, Value: [15]
---Metric: RSRP, Value: [36]
---Metric: RSRQ, Value: [36]

[ersuneetsingh]

Just to add a bit more context about the behavior:

As mentioned earlier, I also tested using ping (one packet per second) with a packet size of 577 bytes. The observed RTT was around 35 ms, but the RLC uplink delay was around 8000 ms, which seems contradictory.

Interestingly, when I reduced the packet size slightly to 566 bytes (i.e., 577 - 1), the RLC uplink delay dropped significantly to just 1 ms.

[newt2105]

hi @ersuneetsingh, what xApp do you use? Can I have your script to run xApp with these metrics?
Thanks!