Configuring QinQ support for SR-IOV networks - Hardware networks | Networking

About 802.1Q-in-802.1Q support
Configuring QinQ support for SR-IOV enabled workloads

QinQ, formally known as 802.1Q-in-802.1Q, is a networking technique defined by IEEE 802.1ad. IEEE 802.1ad extends the IEEE 802.1Q-1998 standard and enriches VLAN capabilities by introducing an additional 802.1Q tag to packets already tagged with 802.1Q. This method is also referred to as VLAN stacking or double VLAN.

About 802.1Q-in-802.1Q support

In traditional VLAN setups, frames typically contain a single VLAN tag, such as VLAN-100, as well as other metadata such as Quality of Service (QoS) bits and protocol information. QinQ introduces a second VLAN tag, where the service provider designates the outer tag for their use, offering them flexibility, while the inner tag remains dedicated to the customer’s VLAN.

QinQ facilitates the creation of nested VLANs by using double VLAN tagging, enabling finer segmentation and isolation of traffic within a network environment. This approach is particularly valuable in service provider networks where you need to deliver VLAN-based services to multiple customers over a common infrastructure, while ensuring separation and isolation of traffic.

The following diagram illustrates how OpenShift Container Platform can use SR-IOV and QinQ to achieve advanced network segmentation and isolation for containerized workloads.

The diagram shows how double VLAN tagging (QinQ) works in a worker node with SR-IOV support. The SR-IOV virtual function (VF) located in the pod namespace, ext0 is configured by the SR-IOV Container Network Interface (CNI) with a VLAN ID and VLAN protocol. This corresponds to the S-tag. Inside the pod, the VLAN CNI creates a subinterface using the primary interface ext0. This subinterface adds an internal VLAN ID using the 802.1Q protocol, which corresponds to the C-tag.

This demonstrates how QinQ enables finer traffic segmentation and isolation within the network. The Ethernet frame structure is detailed on the right, highlighting the inclusion of both VLAN tags, EtherType, IP, TCP, and Payload sections. QinQ facilitates the delivery of VLAN-based services to multiple customers over a shared infrastructure while ensuring traffic separation and isolation.

Diagram showing QinQ (double VLAN tagging)

The OpenShift Container Platform SR-IOV solution already supports setting the VLAN protocol on the SriovNetwork custom resource (CR). The virtual function (VF) can use this protocol to set the VLAN tag, also known as the outer tag. Pods can then use the VLAN CNI plugin to configure the inner tag.

Table 1. Supported network interface cards
NIC	802.1ad/802.1Q	802.1Q/802.1Q
Intel X710	No	Supported
Intel E810	Supported	Supported
Mellanox	No	Supported

Additional resources

Configuration for an VLAN additional network

Configuring QinQ support for SR-IOV enabled workloads

Prerequisites

You have installed the OpenShift CLI (oc).
You have access to the cluster as a user with the cluster-admin role.
You have installed the SR-IOV Network Operator.

Procedure

Create a file named sriovnetpolicy-810-sriov-node-network.yaml by using the following content:

apiVersion: sriovnetwork.openshift.io/v1
kind: SriovNetworkNodePolicy
metadata:
  name: sriovnetpolicy-810
  namespace: openshift-sriov-network-operator
spec:
  deviceType: netdevice
  nicSelector:
    pfNames:
      - ens5f0#0-9
  nodeSelector:
    node-role.kubernetes.io/worker-cnf: ""
  numVfs: 10
  priority: 99
  resourceName: resource810

Create the SriovNetworkNodePolicy object by running the following command:
```
$ oc create -f sriovnetpolicy-810-sriov-node-network.yaml
```
Open a separate terminal window and monitor the synchronization status of the SR-IOV network node state for the node specified in the openshift-sriov-network-operator namespace by running the following command:
```
$ watch -n 1 'oc get sriovnetworknodestates -n openshift-sriov-network-operator <node_name> -o jsonpath="{.status.syncStatus}"'
```
The synchronization status indicates a change from InProgress to Succeeded.
Create a SriovNetwork object, and set the outer VLAN called the S-tag, or Service Tag, as it belongs to the infrastructure.

You must configure the VLAN on the trunk interface of the switch. In addition, you might need to further configure some switches to support QinQ tagging.
1. Create a file named nad-sriovnetwork-1ad-810.yaml by using the following content:
  apiVersion: sriovnetwork.openshift.io/v1 kind: SriovNetwork metadata: name: sriovnetwork-1ad-810 namespace: openshift-sriov-network-operator spec: ipam: '{}' vlan: 171 (1) vlanProto: "802.1ad" (2) networkNamespace: default resourceName: resource810
  1 Sets the S-tag VLAN tag to 171.
  
  2 Specifies the VLAN protocol to assign to the virtual function (VF). Supported values are 802.1ad and 802.1q. The default value is 802.1q.
2. Create the object by running the following command:
  $ oc create -f nad-sriovnetwork-1ad-810.yaml

Create a NetworkAttachmentDefinition object with an inner VLAN. The inner VLAN is often referred to as the C-tag, or Customer Tag, as it belongs to the Network Function:

Create a file named nad-cvlan100.yaml by using the following content:

apiVersion: k8s.cni.cncf.io/v1
kind: NetworkAttachmentDefinition
metadata:
  name: nad-cvlan100
  namespace: default
spec:
  config: '{
    "name": "vlan-100",
    "cniVersion": "0.3.1",
    "type": "vlan",
    "linkInContainer": true,
    "master": "net1", (1)
    "vlanId": 100,
    "ipam": {"type": "static"}
  }'

1	Specifies the VF interface inside the pod. The default name is `net1` as the name is not set in the pod annotation.

Apply the YAML file by running the following command:
```
$ oc apply -f nad-cvlan100.yaml
```

Verification

Verify QinQ is active on the node by following this procedure:

Create a file named test-qinq-pod.yaml by using the following content:

apiVersion: v1
kind: Pod
metadata:
  name: test-pod
  annotations:
    k8s.v1.cni.cncf.io/networks: sriovnetwork-1ad-810, nad-cvlan100
spec:
  containers:
    - name: test-container
      image: quay.io/ocp-edge-qe/cnf-gotests-client:v4.10
      imagePullPolicy: Always
      securityContext:
        privileged: true

Create the test pod by running the following command:
```
$ oc create -f test-qinq-pod.yaml
```

Enter into a debug session on the target node where the pod is present and display information about the network interface ens5f0 by running the following command:

$ oc debug node/my-cluster-node -- bash -c "ip link show ens5f0"

Example output

6: ens5f0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP mode DEFAULT group default qlen 1000
link/ether b4:96:91:a5:22:10 brd ff:ff:ff:ff:ff:ff
vf 0 link/ether a2:81:ba:d0:6f:f3 brd ff:ff:ff:ff:ff:ff, spoof checking on, link-state auto, trust off
vf 1 link/ether 8a:bb:0a:36:f2:ed brd ff:ff:ff:ff:ff:ff, vlan 171, vlan protocol 802.1ad, spoof checking on, link-state auto, trust off
vf 2 link/ether ca:0e:e1:5b:0c:d2 brd ff:ff:ff:ff:ff:ff, spoof checking on, link-state auto, trust off
vf 3 link/ether ee:6c:e2:f5:2c:70 brd ff:ff:ff:ff:ff:ff, spoof checking on, link-state auto, trust off
vf 4 link/ether 0a:d6:b7:66:5e:e8 brd ff:ff:ff:ff:ff:ff, spoof checking on, link-state auto, trust off
vf 5 link/ether da:d5:e7:14:4f:aa brd ff:ff:ff:ff:ff:ff, spoof checking on, link-state auto, trust off
vf 6 link/ether d6:8e:85:75:12:5c brd ff:ff:ff:ff:ff:ff, spoof checking on, link-state auto, trust off
vf 7 link/ether d6:eb:ce:9c:ea:78 brd ff:ff:ff:ff:ff:ff, spoof checking on, link-state auto, trust off
vf 8 link/ether 5e:c5:cc:05:93:3c brd ff:ff:ff:ff:ff:ff, spoof checking on, link-state auto, trust on
vf 9 link/ether a6:5a:7c:1c:2a:16 brd ff:ff:ff:ff:ff:ff, spoof checking on, link-state auto, trust off

The vlan protocol 802.1ad ID in the output indicates that the interface supports VLAN tagging with protocol 802.1ad (QinQ). The VLAN ID is 171.

1	Sets the S-tag VLAN tag to `171`.
2	Specifies the VLAN protocol to assign to the virtual function (VF). Supported values are `802.1ad` and `802.1q`. The default value is `802.1q`.