Virtual Router Redundancy Protocol (VRRP)

Virtual Router Redundancy Protocol (VRRP) is similar to HSRP, but it is an open standard (RFC 2338). Two or more devices act as a virtual router. With VRRP, however, the IP address used can be either a virtual one or the actual IP address of the primary router.

The VRRP Master router forwards traffic. The master is chosen because 1) it owns the real address, or 2) it has the highest priority (default is 100). If a real address is being supported, the owner of real address must be master. A Backup router takes over if the master fails, and there can be multiple backup routers. They monitor periodic hellos multicast by the master to 224.0.0.18, using UDP port 112, to detect a failure of the master router.

Multiple VRRP groups are allowed, just as with HSRP.

Routers in the same VRRP group must belong to the same subnet/VLAN. To enable VRRP, give this command vrrp groupnumber ip virtual-IP-address under the interface connecting to that
subnet or VLAN:

Router(config-if)#vrrp 39 ip 10.0.0.1

Control the master and backup elections by configuring priority values from 1–255. If a master VRRP router is shutdown, it advertises a priority of 0. This triggers the backup routers to hold an election without waiting for the master’s hellos to time out.

Router(config-if)#vrrp 39 priority 175

VRRP uses the following timers:

• Advertisement, or hello, interval in seconds. Default is 1 second.

• Master down interval. Equals (3 x advertisement interval) plus skew time. Similar to a hold or dead timer.

• Skew time. (256–priority) / 256. This is meant to ensure that the highest priority backup router becomes master, since higher priority routers have shorter master down intervals.

To change the timers on the master, use the following command because it is the router that advertises the hellos:

Router(config-if)#vrrp 39 timers advertise 5

To change the timers on the backup routers, use the following command because they hear the hellos from the master:

Router(config-if)#vrrp 39 timers learn


GLBP

One issue with both HSRP and VRRP is that only the primary router is in use, the others must wait for the primary to fail before they are used. These two protocols use groups to get around that limitation. However, Gateway Load Balancing Protocol (GLBP) allows the simultaneous use of up to four gateways, thus maximizing bandwidth. With GLBP, there is still one virtual IP address. However, each participating router has a virtual MAC address, and different routers’ virtual MAC addresses are sent in answer to ARPs sent to the virtual IP address. GLBP can also use groups up to a maximum of 1024 per physical interface.

The load sharing is done in one of three ways:

• Weighted load balancing—Traffic is balanced proportional to a configured weight.

• Host-dependent load balancing—A given host always uses the same router.

• Round-robin load balancing—Each router MAC is used to respond to ARP requests in turn.

GLBP routers elect an Active Virtual Gateway (AVG). It is the only router to respond to ARPs. It uses this capacity to balance the load among the GLBP routers. The highest priority router is the AVG; the highest configured IP address is used in case of a tie.

The actual router used by a host is its Active Virtual Forwarder (AVF). GLBP group members multicast hellos every 3 seconds to IP address 224.0.0.102, UDP port 3222. If one router goes down, another router answers for its MAC address.

Configure GLBP with the interface command glbp group-number ip virtual-IP-address, as shown:

Router(config-if)#glbp 39 ip 10.0.0.1

To ensure deterministic elections, each router can be configured with a priority. The default priority is 100:

Router(config-if)#glbp 39 priority 150


Hello and hold (or dead) timers can be configured for each interface with the command glbp group-number timers [msec] hello-time [msec] hold-time. Values are in seconds unless the msec keyword is used.

GLBP can also track interfaces; if an interface goes down, another router answers for the first router’s MAC address.