What is a Routing Instance?

A routing instance is a named, reusable forwarding context on a Versa VOS device – its own routing table and interface bindings – that the device consults to reach a destination. Administrators define it once, then reference it by name when configuring interfaces, static routes, monitors, DHCP, and SaaS application monitor profiles.

Key takeaways

• A routing instance is a named forwarding context on a VOS device.
• Created once, then referenced by name across many configuration screens.
• Every device has a global routing instance, reserved ID 0, automatically.
• LAN interfaces, static routes, HA links, and monitors all select one.

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

Why are routing instances important?

A VOS device typically needs several forwarding contexts at once: one tenant’s LAN traffic, another’s WAN transport, a path for HA synchronization, and a global context for traffic belonging to no tenant. Defining that context once, under a name, and referencing it everywhere keeps configuration consistent and avoids re-specifying a routing table and interface bindings on every feature that forwards traffic.

That reuse is what makes routing instances load-bearing across the platform. A LAN interface, a static route, an HA link, a multicast deployment, a DHCP relay, and a SaaS monitor all ask the same question: which routing instance does this operate in? Getting it wrong routes traffic into the wrong context, or drops it, since traffic outside a tenant’s routing instance is discarded rather than rerouted.

What is a routing instance?

In Director, a routing instance is a configuration object you name once, then select wherever a feature needs a forwarding context. Creating one takes a name, an optional description, a numeric ID, and, if it terminates VPN traffic, a VPN option – see Configure Basic Features, which exposes a Routing Instances tab in both the classic Organizations form and the multi-step workflow.

Every routing instance maintains its own routing table and interface bindings; an interface, subinterface, or loopback belongs to exactly one at a time. Traffic not assigned to a routing instance lands in the global routing instance, reserved ID 0, present on every device automatically – but traffic outside a tenant is dropped rather than routed through it. Monitoring commands take a routing instance as a parameter, scoping diagnostics to the exact context under investigation.

Types of routing instances

Every VOS device has a global routing instance, reserved ID 0, created automatically for unassigned traffic. Beyond that, routing instances are created per organization on the Routing Instances tab, each with a name, ID, and optional VPN flag.

Several roles recur by convention rather than a fixed type field. A control VR routing instance carries the IPsec tunnels and control-plane traffic between an organization and the Director and Controller nodes – an organization either keeps its own or shares its parent’s under a shared control plane. LAN and transport routing instances separate a tenant’s internal routes from WAN-facing underlay routes. Interchassis HA adds a dedicated pair: a data routing instance (often a purpose-built virtual router, such as HA-Control-VR) for the stateful sync connection, and a separate control routing instance for peer discovery, configurable over a LAN interface or the management interface instead.

Routing Instances & Versa

Across Director workflows, routing instance selection shows up wherever traffic needs a forwarding context. Creating or modifying an organization includes a Routing Instances tab, and device templates ask for one on every LAN interface and static route – see Configure Basic Features.

Multicast deployments scope PIM to a routing instance explicitly: an SD-WAN multicast cluster is defined by tenant ID plus routing instance, and diagnostic commands such as show pim interface require one as a parameter – see Configure IP Multicast.

Interchassis HA depends on routing instances to keep its own traffic separate from tenant traffic: the data and control TCP connections can each be pinned to a specific routing instance, falling back to the global one only when neither is configured – see Configure Stateful Interchassis HA. DHCP relay and SaaS monitor profiles reference routing instances the same way, by name.

Frequently asked questions

What is the difference between a routing instance and a VRF?

Same underlying mechanism: a routing instance is Versa’s configuration object for VRF-style routing separation – its own table, its own bound interfaces. “Virtual Router” (VR) further names specific instances, such as the LAN VR, transport VR, or control VR, filling particular roles within an organization’s routing instances.

What is the global routing instance?

It’s the routing instance with reserved ID 0, created automatically on every VOS device. Interfaces not explicitly assigned elsewhere land there by default, but traffic outside a tenant is dropped rather than forwarded through it, so it isn’t a general-purpose fallback.

Can two different features reference the same routing instance?

Yes – that’s the reason to name it once. A single routing instance can serve as a LAN interface’s forwarding context, the routing table for its static routes, and the scope for its DHCP relay and SaaS monitor profile, all at once.

Why does interchassis HA use its own routing instance?

To keep synchronization traffic isolated from tenant data. Interchassis HA uses one routing instance for the stateful data sync connection and a separate control routing instance for peer discovery, so HA traffic never competes with or leaks into tenant routing tables.

How many routing instances can one organization have?

The documentation doesn’t set one fixed cap; an organization defines as many named routing instances as its deployment needs on the Routing Instances tab, each with its own ID, in addition to the automatic global routing instance every device already has.

More on routing instances

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

Related terms

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