As I know, VXLAN uses OSPF for its underlay. However, it seems BGP is involved in the underlay, but I cannot figure out the purpose.
A VTEP encapsulates a packet, and if OSPF is deployed, every device should be able to route the packet to another VTEP via the outer-MAC and outer-IP. What BGP actually does in the VXLAN?
Apparently you have to have BGP.
A VLAN is configured for each segment - sending segment, VRF segment and receiving segment.
BGP and EVPN configurations ensure redistribution of this information across the VXLAN setup.https://www.cisco.com/c/en/us/td/docs/switches/datacenter/pf/configuration/guide/b-pf-configuration/Introducing-Cisco-Programmable-Fabric-VXLAN-EVPN.htmlHTH
Thanks for your information. I just finished it. There are another questions:
1.) If VTEPs advertises every end host’s information while they attach, in what circumstance ARP request is needed? Will the information timeout?
2.) As I know, EVPN seems to be related to MPLS. I would like to ask how SPINEs route the packets? Do they parse the outer-IP and look up their routing table?
If there are 10 vteps(switches), how could VTEP1 know to send packet through Vxlan to VTEP8, and not VTEP6? It looks to BGP table where the table of Mac, VTEP pairings exist