What is IBGP?

Internal BGP (IBGP) is a BGP session established between routers that belong to the same autonomous system (AS). It carries routes learned from external BGP (EBGP) peers to every router inside the AS, keeping the network synchronized on the same externally learned routing information.

Key takeaways

• IBGP distributes externally learned routes among routers in one AS.
• A loop-prevention rule blocks re-advertising IBGP routes to other IBGP peers.
• Route reflectors let IBGP scale without a full peer mesh design.
• VOS gives IBGP routes a default preference of 200, versus 20 for EBGP.
• Versa’s SD-WAN control plane defaults to IBGP AS number 64512.

On this page: Why it matters · How it works · Types · IBGP & Versa · FAQ

Why is IBGP important?

Any AS that connects to the outside world through more than one router needs a way to share what those border routers learn. IBGP is that mechanism: it lets every router inside the AS see the same external routes, so traffic entering at one edge is forwarded toward an exit managed by a different router. Without it, only the router that directly received a route from an EBGP peer would know about it.

IBGP also keeps the AS acting as one coherent routing domain externally. Attributes learned over EBGP – AS path, communities, local preference – carry through unchanged across IBGP sessions, so policy decisions made at one edge stay consistent everywhere. That is what lets a multi-router AS behave as one predictable neighbor to the rest of the internet.

What is IBGP?

An IBGP session is a BGP session where both peers configure the same AS number, unlike EBGP, where the peers use different AS numbers. IBGP peers are usually not directly connected, so the session typically runs over an interior gateway protocol or static routes that provide reachability between the peers’ addresses.

The defining rule of IBGP is loop prevention: a router must not re-advertise a route learned from one IBGP peer to another. This keeps IBGP routes from looping inside the AS, but it also means every router needs a direct session with every other router – a full mesh – unless that requirement is managed. That is the main practical difference from EBGP, which has no equivalent restriction.

Types of IBGP

Three approaches solve the IBGP full-mesh scaling problem. Full mesh is the baseline: every router peers directly with every other router – no extra design, but it grows as n(n-1)/2 sessions, unmanageable past a few dozen routers.

Route reflection lets a designated router – a route reflector – accept routes from its IBGP clients and re-advertise them to other clients, bypassing the loop-prevention rule for that relationship.

Confederations split a large AS into smaller sub-ASes that peer via EBGP internally while appearing as one AS externally – avoiding the full-mesh problem but adding more complexity than route reflection.

IBGP & Versa

In VOS, IBGP is a first-class BGP session type distinct from EBGP, with its own defaults. A peer group’s Type field is set to EBGP or IBGP, and VOS gives each a different route preference: IBGP-learned routes default to 200, EBGP-learned routes to 20. On TVI-based WAN interfaces, VOS enables Next-Hop Self by default on IBGP sessions, so the local router advertises itself as next hop rather than an address other AS members may not reach. Route Reflector Client is an IBGP-only setting, letting a device act as a reflector instead of requiring a full mesh; the full field set is in Configure BGP.

Versa’s SD-WAN control plane relies heavily on IBGP. Per the SD-WAN design guide, Hub-Controller Nodes establish IBGP sessions with the Head-End Controller to unify routing in restricted or isolated regions, defaulting to AS number 64512. IBGP also runs between redundant edge devices over a cross-connect link, giving a backup device full route visibility instead of relying on a default route.

Frequently asked questions

What is the difference between IBGP and EBGP?

IBGP connects peers in the same AS and distributes externally learned routes internally; EBGP connects peers in different ASes and is how those routes get learned. VOS applies different defaults to each – EBGP-learned routes get a preference of 20, IBGP-learned routes get 200.

Why does IBGP require a full mesh or route reflectors?

IBGP’s loop-prevention rule stops a router from re-advertising a route learned from one IBGP peer to another IBGP peer. Without a full mesh, some routers would never receive certain routes; route reflectors or confederations solve this without meshing every pair of routers.

What does Next-Hop Self do on an IBGP session?

Next-Hop Self tells a router to advertise itself as the next hop instead of preserving the original EBGP next hop, which other routers inside the AS may not be able to reach. VOS enables this by default on IBGP sessions configured on TVI-based WAN interfaces.

What is a route reflector?

A route reflector is an IBGP router that re-advertises routes received from one IBGP client to its other IBGP clients, bypassing the standard loop-prevention rule for that relationship. VOS supports this as the Route Reflector Client setting on an IBGP peer group.

What AS number does Versa’s SD-WAN control plane use for IBGP?

Versa’s SD-WAN control plane defaults to AS number 64512 for the internal IBGP sessions that unify routing between Hub-Controller Nodes and the Head-End Controller. This matters most in restricted or isolated regions, where each region may run its own controller instance and still needs a single, coherent control plane.

More on IBGP

If you want to learn more, these are sources we trust.

Related terms

If you’re reading about IBGP, these are worth looking at next.