VRFs and route-leaking

 Route leaking is the controlled process of sharing specific network routing prefixes between otherwise isolated Virtual Routing and Forwarding (VRF) instances or between a VRF and the Global Routing Table (GRT).

Why Use Route Leaking?

Shared Services: Allow multiple isolated customer or department VRFs to access common resources like DNS, DHCP, or firewalls.
Internet Breakout: Route traffic from a private VRF out to a shared public gateway or the global routing table.
Multi-Tenant Communication: Enable authorized inter-tenant data flows without merging entire routing tables into a single insecure domain. 

I want to connect SOME blocks. I want Blue, Green, and Yellow to be able to communicate with the “Core” off the GRT (global routing table) VRF. I do not want Blue, Green, or Yellow to be able to communicate with each other though.

However, what if you have some shared services or routes that should be shared between multiple VRFs?

It is possible to “leak” routes from one VRF into another. There are 3 options to achieve this:

Common Methods for Route Leaking

• MP-BGP (Multi-Protocol BGP): Uses route target (RT) export and import statements to exchange routing information between VRFs via a local VPN table.
Static Routes: Configures direct static routes that point to an egress interface or next-hop IP belonging to a different routing context (supported natively in modern platforms like FortiOS or via GRT integration).
Route Maps & Agents: Leverages policy-based controls, such as Arista's VRF-leak agent or Cisco's route replication, to filter which specific prefixes cross the boundary.


Here is another example with different networks