Simple definition
A VRF (Virtual Routing and Forwarding) keeps multiple isolated routing tables on the same router or Layer 3 device.
Technical definition
Each VRF has its own routing context and can use separate interfaces, static routes, and dynamic routing protocols. The same IP prefix can therefore exist in different VRFs without directly conflicting.
How it works / role
An interface is assigned to a VRF, and packets received on that interface are looked up in that VRF’s routing table. Traffic reaches another VRF only when an explicit route-leaking or interconnection method is configured.
What is it used for?
Isolate customers, environments, services, or network segments without deploying a separate physical router for each one.
Practical example
A provider hosts two customers that both use 10.0.0.0/24 in separate VRFs and keeps their routes independent.
Common issues
- Interface is assigned to the wrong VRF
- Route exists in the global table but not in the VRF
- Route leaking is missing or too permissive
- Troubleshooting is performed from the wrong routing context
Key takeaway: A VRF primarily isolates routing decisions; communication between VRFs must be designed explicitly.