What Is VRRP? Virtual Router Redundancy Protocol Explained

VRRP (Virtual Router Redundancy Protocol) is a standard protocol that lets two or more routers share one virtual default gateway address, so a backup router takes over automatically if the master fails.

Quick answer: VRRP (Virtual Router Redundancy Protocol) is a standard protocol that lets two or more routers share one virtual default gateway address, so a backup router takes over automatically if the master fails.

VRRP (Virtual Router Redundancy Protocol) is a standard networking protocol that lets two or more routers share one virtual default gateway address on a LAN. One router, the master, answers for that address. If it fails, a backup router takes the address over automatically, so the devices behind it keep the same gateway setting and never need reconfiguring.

I think of VRRP as the bit of plumbing that turns “we’ve got a spare router in the cabinet” into “the spare router actually takes over”. Without it, every PLC, RTU and camera on the LAN points at one router’s address, and when that router dies, someone has to drive out and change things. With it, the devices point at an address that belongs to whichever router is currently working.

How VRRP works

The routers in a VRRP group share a virtual router ID (VRID) and one or more virtual IP addresses. Each router has a priority from 1 to 254, with 100 as the usual default. The router with the highest priority becomes master and answers for the virtual address. The others sit as backups and listen.

The master proves it’s alive by sending advertisements on the LAN. In VRRP version 3, these are multicast to 224.0.0.18 (or FF02::12 for IPv6) using IP protocol 112, with a TTL of 255 so they never leave the local network. The default advertisement interval is one second. The standard expresses this in centiseconds, which is why some router menus show a default of 100.

If the backups stop hearing advertisements, each waits for its Master_Down_Interval: three advertisement intervals plus a small skew time based on its priority. The higher-priority backup times out first and takes over. With default settings, that’s a little over three seconds.

TermWhat it means
Virtual router ID (VRID)A number from 1 to 255 that groups routers into one virtual router. Every router in the group must use the same one.
Virtual IPThe shared gateway address devices point at.
MasterThe router currently answering for the virtual IP.
BackupA router waiting to take over if the master goes quiet.
Priority1 to 254 decides who’s master. 255 is reserved for a router that actually owns the virtual IP as its real address. A master sending priority 0 is announcing it’s stepping down, so a backup can take over straight away.
PreemptionWhether a higher-priority router takes the master role back when it returns. On by default in the standard.
Virtual MACA shared MAC address, 00-00-5E-00-01-{VRID} for IPv4, so devices don’t need to relearn the gateway’s hardware address after a handover.

VRRP versions and standards

VRRP is an open IETF standard, which is a big part of its appeal: routers from different vendors can, in principle, share a VRRP group. Version 2 (RFC 3768) was IPv4 only. Version 3 (RFC 5798) added IPv6 and moved to centisecond timers. RFC 9568, published in April 2024, is the current VRRPv3 specification and replaces RFC 5798.

One detail people get wrong: VRRPv3 has no authentication. Earlier versions had authentication options, but the standard dropped them as giving a false sense of security. Anything on the LAN segment could, in theory, send advertisements, which is another reason to keep OT networks segmented.

Cisco’s HSRP and OpenBSD’s CARP do a similar job, but they’re not VRRP and won’t interoperate with it.

What VRRP doesn’t do on its own

This is the part that matters for cellular IoT. Plain VRRP only notices when the master router stops sending advertisements, in other words when it’s switched off, crashed or unplugged from the LAN. It knows nothing about the router’s internet connection.

That’s a problem for a 4G router. The commonest failure isn’t the router dying. It’s the router carrying on perfectly happily on the LAN while its mobile data path has gone nowhere. Plain VRRP will leave that router as master all day.

So most router vendors add some form of tracking. The router monitors something beyond the LAN, a ping to an internet host or the state of a WAN interface, and steps down or lowers its priority when that check fails. On Teltonika’s RUTOS routers, for example, VRRP is an add-on package with an optional connection check that pings a host you choose and hands the master role to a backup when the replies stop.

The trap with ping-based tracking: the check follows the router’s own routing table. If the master has any route out through the backup router, its pings can succeed via the backup and it will hang onto the gateway with a dead uplink. And a ping target can answer while the platform your equipment actually reports to is unreachable. Tracking tells you the router can reach one host. It doesn’t prove your data is getting through.

Where VRRP is used in IoT

  • Two-router cellular resilience. Two 4G or 5G routers on different mobile networks share one gateway address, so a remote cabinet survives the loss of a router or a network. I’ve gone through this design in detail, including all the ways it can catch you out, in my piece on bonded 4G versus a second router.
  • Fixed line with cellular backup in a separate box. The fixed-line router is master, a 4G router is backup, and LAN devices keep one gateway setting.
  • Industrial and utility sites where a single gateway is an unacceptable single point of failure for SCADA traffic.

The alternative is a single dual-WAN or dual-modem router doing failover internally. That’s simpler, but it’s still one box. VRRP is how you get two boxes to act as one gateway.

Common VRRP problems in the field

ProblemWhat causes it
Master keeps the gateway with a dead uplinkNo tracking, or tracking that can reach its target another way.
Gateway flaps back and forthA marginal link combined with aggressive check settings or preemption.
Devices keep sending to the old routerThe device caches the old MAC address and ignores gratuitous ARP. A virtual MAC usually fixes it.
Both routers think they’re masterAdvertisements not reaching each other, often a VLAN, switch or firewall issue.
Original master doesn’t take back overPreemption off, a tracking state that never clears, or a firmware bug. Test failback after every update.
DHCP clients get the wrong gatewayThe DHCP server hands out the router’s own address instead of the virtual IP.

My advice is always the same: don’t trust any of it until you’ve watched it fail over and back on the firmware you’re actually deploying, with the real equipment plugged in.

Quick answers

What does VRRP stand for?

Virtual Router Redundancy Protocol. It’s an IETF standard that lets several routers share one virtual default gateway address.

How fast is VRRP failover?

With default timers, a backup takes over a little over three seconds after the master stops advertising. Failover triggered by connection tracking also includes the time the check takes to declare the uplink dead, which depends on its interval and retry settings.

What’s the difference between VRRP and HSRP?

Both provide a shared gateway between routers. VRRP is an open standard supported by many vendors. HSRP is Cisco’s proprietary protocol.

Does VRRP detect a failed internet connection?

Not on its own. Standard VRRP only detects a master that stops advertising on the LAN. Vendors add tracking, such as a ping check or WAN interface monitoring, to trigger failover when the internet connection fails.

Can routers from different vendors use VRRP together?

In principle yes, because it’s a standard, provided they run the same VRRP version with matching IDs and timers. Vendor-specific tracking features won’t carry across, so test mixed pairs carefully.