No relevant resource is found in the selected language.

This site uses cookies. By continuing to browse the site you are agreeing to our use of cookies. Read our privacy policy>Search


To have a better experience, please upgrade your IE browser.


Guide for Interworking Between HUAWEI CloudFabric Solution and Redhat OpenStack

Rate and give feedback :
Huawei uses machine translation combined with human proofreading to translate this document to different languages in order to help you better understand the content of this document. Note: Even the most advanced machine translation cannot match the quality of professional translators. Huawei shall not bear any responsibility for translation accuracy and it is recommended that you refer to the English document (a link for which has been provided).
Distributed Network Overlay

Distributed Network Overlay

Introduction to the Physical Baseline Network

The underlay network uses the mature spine-leaf architecture to provide the scale-out capability and high access performance in the area. Leaf and spine nodes are fully meshed so that ECMP paths are available to ensure high availability of the network.

  • Spine node: In most cases, a spine node is composed of two large-capacity switches, and the Ethernet interfaces connected to leaf nodes are configured as routed interfaces to establish an IP fabric network.
  • Leaf node: Leaf nodes are connected to each spine node. Leaf nodes are the L2/L3 boundary on the underlay network, and their Ethernet interfaces connected to the spine node are configured as routed interfaces. Because many ToR switches need to be deployed at leaf nodes, the Zero Touch Provisioning (ZTP) mode is recommended to simplify the deployment.
    Figure 2-1 Spine-leaf architecture
  • Servers can be dual-homed to server leaf nodes through M-LAG, as shown in Figure 2-2.
    Figure 2-2 Servers dual-homed to server leaf nodes through M-LAG

The following table lists the recommended models.

Device Role

Recommended Model

Server Leaf


Service Leaf


Boarder Leaf




Fabric Gateway


Introduction to the Overlay

Figure 2-3 shows the typical network overlay networking.

Figure 2-3 Typical network overlay networking
  • Control plane principles
    1. The Agile Controller-DCN uses NETCONF interfaces to deploy IBGP EVPN of the overlay on each leaf node.
    2. The spine node functions as an RR, and each NVE node functions as a client.
    3. BGP EVPN uses inclusive routes (Type 3 routes) to transmit Layer 2 VNIs and VTEP addresses, dynamically creating VXLAN tunnels and ingress replication tables and preventing fully-meshed static tunnel configuration.
  • Entry synchronization on the control plane
    1. NVE nodes learn forwarding entries such as MAC or ARP entries from data packets or ARP packets sent from VMs.
    2. The NVE node imports learned forwarding entries to EVPN instances and forms EVPN protocol routes.
    3. The NVE node advertises routes (Type-2 MAC/IP routes) to its peers through BGP EVPN.
    4. The peers (NVE nodes) generate forwarding entries after receiving MAC or IP routes.
    5. BGP EVPN advertises MAC or IP routes (Type-2) and IP prefix routes (Type-5) to ensure connectivity of Layer 2 and Layer 3 forwarding paths.
Updated: 2019-03-25

Document ID: EDOC1100072313

Views: 2360

Downloads: 15

Average rating:
This Document Applies to these Products
Related Documents
Related Version
Previous Next