Lab 10: Slicing rApp for NTN

Experimental

This is an experimental blueprint intended for Non-Terrestrial Network (NTN) evaluation in an emulated environment.

In this lab, you will deploy a 5G Standalone (SA) Non-Terrestrial Network (NTN). The network consists of an OAI NTN gNB, an OAI UE, the OAI minimal 5G Core (5GC), and FlexRIC as the Near-RT RIC. The satellite link between the UE and the gNB is emulated with the OAI RF Simulator in a Low Earth Orbit (LEO) scenario.

In the second part of the lab, you will use the BubbleRAN rApp SDK to run a sample rApp that sends a slicing policy to the Non-RT RIC over the R1 interface. The Non-RT RIC forwards the policy to the Near-RT RIC over the A1 interface. An xApp then enforces the policy at the gNB over the E2 interface. The result is a network slice whose downlink Physical Resource Block (PRB) usage is capped at the NTN gNB.

Network Description

The network is defined in the following file:

ntn-oai-net.yaml

apiVersion: athena.trirematics.io/v1
kind: Network
metadata:
    name: oai-rfsim-ntn-leo
    namespace: trirematics
spec:
    slices:
        -   plmn: "00101"
            dnn: internet
            network-mode: IPv4
            service-type: 1
            differentiator: 0x000000
            ipv4-range: 12.1.1.0/24
            ipv6-range: 2001:db8:1::/64
    access:
        -   name: gnb
            stack: 5g-sa
            model: oai-ran/monolithic-gnb-ntn
            identity:
                an-id: 30
                tracking-area: 1
            radio:
                device: rf-sim
            annotations:
                extras.t9s.io/ntn-scenario: "leo"
            cells:
                -   band: n254
                    arfcn: 497680
                    bandwidth: 5MHz
                    subcarrier-spacing: 15kHz
            core-networks:
                - minimal.oai-rfsim-ntn-leo
            controller: flexric.oai-rfsim-ntn-leo
    edge:
        -   name: flexric
            stack: 5g-sa
            model: mosaic5g/flexric
    core:
        -   name: minimal
            stack: 5g-sa
            model: oai-cn/minimal
            identity:
                region: 0
                cn-group: 4
                cn-id: 5
    dns:
        ipv4:
            default: 8.8.8.8
            secondary: 8.8.4.4
---
apiVersion: athena.trirematics.io/v1
kind: Terminal
metadata:
    name: oai-ue
    namespace: trirematics
    annotations:
        extras.t9s.io/ntn-scenario: "leo"
spec:
    vendor: oai
    stack: 5g-sa
    model: terminal/oai-ue-ntn
    preferred-access: gnb.oai-rfsim-ntn-leo
    target-cores:
        - minimal.oai-rfsim-ntn-leo
    identity:
        imsi: "001010000000001"
        pin: "1234"
        opc: "0xc42449363bbad02b66d16bc975d77cc1"
        key: "0xfec86ba6eb707ed08905757b1bb44b8f"
        sqn: "0xff9bb4000001"
    slice:
        dnn: internet
        network-mode: IPv4
        service-type: 1
        differentiator: 0x000000
    radio:
        device: rf-sim
        formation: "1x1"
        bands:
            - n254
    readiness-check:
        method: ping
        target: google-ip
        interface-name: oaitun_ue1

Compared with a terrestrial deployment, note the following NTN-specific settings:

• The gNB uses the oai-ran/monolithic-gnb-ntn model, and the UE uses the terminal/oai-ue-ntn model.
• The cell operates in band n254, an NTN band.
• The extras.t9s.io/ntn-scenario: "leo" annotation, set on both the gNB and the UE, selects the satellite orbit. It must be the same on both sides.

Network Deployment

Deploy the network:

brc install network ntn-oai-net.yaml

tip

The ntn-oai-net.yaml file does not include an xApp. The SMO deploys the xApp automatically later in the lab, when the rApp sends its policy.

Run brc observe to check the deployment, and wait until all Elements are in the 1/1 Y state.

Before applying any slice policy, measure the UE's downlink throughput to get a baseline:

brc test throughput oai-ue dl gateway -- -t 60

Write down the result. You will compare it with the throughput after slice enforcement.

Slice Enforcement via rApp

The following rApp applies a policy named slice1 to the Near-RT RIC flexric.oai-rfsim-ntn-leo:

slicing-rapp-ntn.py

from br_rapp_sdk import A1Services
from br_rapp_sdk.a1_services.a1_policy_types import *


if __name__ == "__main__":

    # Initialize the A1Services
    a1_services = A1Services()

    # Define the policy information
    policy_info = PolicyObjectInformation(
        near_rt_ric_id =NearRtRicId("flexric.oai-rfsim-ntn-leo"),
        policy_type_id=PolicyTypeId("bubbleran/slicing"),
        policy_object=PolicyObject(
            scope_identifier=ScopeIdentifier(
                slice_id=SliceId(
                    sst=1,
                    sd="000000"
                ),
                cell_id=CellId(
                    plmn_id=PlmnId(
                        mcc="001",
                        mnc="01"
                    )
                ),
            ),
            policy_statements=PolicyStatements(
                policy_objectives=PolicyObjectives(
                    lb_objectives=LbObjectives(
                        target_prb_usg=10,
                        prb_usg_type=7
                    )
                )
            )
        )
    )

    # Print the policy information in YAML format
    print("Policy Object Information: \n", policy_info.yaml())

    # Apply the policy
    policy_name= "slice1"
    result = a1_services.apply_policy(policy_name = policy_name, policy_object = policy_info)

    # Check if the operation was successful
    if result.status == 'success':
        policy_id = result.data.get('policy_id')
        print(f"Policy applied successfully: {policy_id}")
    else:
        print(f"Error applying policy: {result.error}")

The policy works as follows:

• Policy type: bubbleran/slicing.
• Scope: the slice with SST 1 and SD 000000 in PLMN 00101. This is the slice the UE is attached to.
• Statement: a load-balancing objective with target_prb_usg=10, which limits the slice to 10% of the downlink PRBs.

danger

Before you run the rApp:

• Make sure the BubbleRAN rApp SDK is installed. If it is not, install it with pip3 install br-rapp-sdk.
• Run the rApp inside the cluster, and make sure your kubeconfig file is located at ~/.kube/config.

Run the rApp:

python3 ./slicing-rapp-ntn.py

The rApp prints the policy in YAML format, followed by Policy applied successfully and the policy ID.

Run brc observe again. A new xApp named slicing.dynxapp-flexric.oai-rfsim-ntn-leo should appear in the list of deployed xApps. Wait until it is in the 1/1 Y state, which means it is deployed and running.

Measure the UE's downlink throughput again, now with the slice policy enforced:

brc test throughput oai-ue dl gateway -- -t 60

The throughput should be noticeably lower than the baseline, because the slice can now use only 10% of the downlink PRBs.

Uninstall

Remove the network and the policy job:

brc remove network ntn-oai-net.yaml
brc remove policyjob slice1

Run brc observe to confirm that all Elements and rApp jobs have been removed.

💬 Questions

1. What does the LEO scenario mean in this NTN deployment? How does it differ from a GEO scenario in terms of satellite altitude, propagation delay, and Doppler shift?
2. Repeat the lab with the GEO scenario by setting extras.t9s.io/ntn-scenario to "geo" on both the gNB and the UE. Compare the throughput results with those of the LEO scenario, with and without slice enforcement.
3. Extend slicing-rapp-ntn.py to support multiple slices with different PRB limits. Compare the throughput results with the single-slice case.
4. Use the KPM monitoring rApp to monitor the UE's PRB usage and throughput. Compare these values with the throughput reported by brc test throughput.