User's Manual

65-1
Chapter 65 IPv6 VRRPv3 Configuration
65.1 Introduction to VRRPv3
VRRPv3 is a virtual router redundancy protocol for IPv6. It is designed based on VRRP (VRRPv2) in IPv4
environment. The following is a brief introduction to it.
In a network based on TCP/IP protocol, in order to guarantee the communication between the devices which
are not physically connected, routers should be specified. At present there are two most commonly used
methods to specify routers: one is to study dynamically via routing protocols (such as internal routing
protocols RIP and OSPF); the other is to configure statically. Running dynamical routing protocol on each
terminal is unrealistic, since most operating systems for client end do not support dynamical routing protocol,
even if they do, they are limited by the overheads of management, convergence, security and many other
problems. So the common method is to adopt static routing configuration on terminal IP devices, which
usually means specify one or more default gateway for terminal devices. Static routing simplifies the
management of network and reduces the communication overheads of terminal devices, but it still has a
disadvantage: if the router acting as the default gateway breaks, the communication of all the hosts which use
this gateway as their next hop host. Even if there are more than one default gateways, before rebooting the
terminal devices, they can not switch to the new gateway. Adopting virtual router redundancy protocol (VRPR)
can effectively avoid the flaws of statically specifying gateways.
In VRRP protocol, there are two groups of import concepts: VRRP routers and virtual routers, master routers
and backup routers. VRRP routers are routers running VRRP, which are physical entities; virtual routers are
the ones created by VRRP, which are logical concepts. A group of VRRP routers cooperate to comprise a
virtual router, which acts outwardly as a logical router with a unique fixed IP address and MAC address. The
routers belonging to the same VRRP group play two mutually exclusive roles at the same time: master routers
and backup routers. One VRRP group can only have one master router other but one or more backup routers.
VRRPv3 protocol uses selection policy to select a master router from the router group to take charge of
responding ND(Neighbor Discovery) neighbor request messages(ARP in IPv4) and forwarding IP data
packets, while the other routers in the group will be in a state of waiting as backups. When the master router
has a problem for some season, the backup router will be updated to the master router after a delay of a few
seconds. Since this switch is very fast and does not need to change IP address or MAC address, it will be
transparent to terminal user systems.
In IPv6 environment, the hosts in a LAN usually learn the default gateway via neighbor discovery protocol
(NDP), which is implemented based on regularly receiving advertisement messages from routers. The NDP of
IPv6 has a mechanism called Neighbor Unreachability Detection, which checks whether a neighbor node is
failed by sending unicast neighbor request messages to it. In order to reduce the overheads of sending
neighbor request messages, these messages are only sent to those neighbor nodes which are sending flows,
and are only sent if there is no instruction of UP state of the router in a period of time. In Neighbor
Unreachability Detection, if adopting default parameters, it will take about 38 seconds to detect an
unreachable router, which is a delay not ignorable for users and might cause a time-out in some transport