Cisco SDWAN - Why the DN bit is used as loop prevention - OMP vs OSPF

[Meddane]

While investigating and Observing OSPF behavior with SD-WAN and why the DN bit is used as loop prevention mechanism when redistributing between OMP and OSPF. It seems that with Cisco SD-WAN, the WAN Edge which is natively a CE device that connects to MPLS Provider BEHAVES also like a PE Device in SD-WAN Fabric.

And the SD-WAN Fabric behaves like MPLS Superbackbone.

This implies that when running OSPF on service VPN (LAN), putting an ABR in the internal network in the middle between two WAN Edges can cause somes routes ignored by the WAN Edges if they are learned through a non-backbone area.

Also Knowing that originally the DN bit is used in MPLS L3 VPN between PE routers to prevent loops, I think this is why the DN bit is also used as a loop prevention between WAN Edges when redistributing between OMP and OSPF.

[Mohamed Alhenawy]

Yes, you are correct , In SD-WAN, the WAN Edge devices serve as the customer edge (CE) devices connecting to the MPLS provider network. However, in the SD-WAN fabric, these WAN Edge devices also behave as provider edge (PE) devices. The SD-WAN fabric itself acts as a super backbone, similar to the MPLS backbone so When running OSPF on the LAN (service VPN 1-511), if an Area Border Router (ABR) is placed within the internal network between two WAN Edge devices, there is a possibility that routes learned through a non-backbone area may be ignored by the WAN Edges. This behavior is due to the design of the SD-WAN fabric, which treats the internal network like an MPLS VPN network, where routes learned from non-backbone areas are not propagated to the other edges.

[Martin L]

thanks for sharing!