Avaya ERS 8600 series, Configuration Manual

Page: 71 / 420

At startup, R1 and R2 use the IST link to exchange full configuration information including MAC
address for the IPv6 interfaces residing on SMLT VLAN 3.
When R2 detects that the RSMLT in R1 transitions to the DOWN state (for example, if R1 itself
is down, or its SMLT links are down, or the IST link is down) R2 takes over IPv6 termination
and IPv6 Neighbor Discovery functionality on behalf or R1’s IPv6 SMLT interface. Specifically:
• When the above event is detected, R2 transmits an unsolicited IPv6 Neighbor
Advertisement for each IPv6 address configured on R1’s SMLT link using R1’s MAC
address (fe80::1 and 2003::1 in this example).
• R2 also transmits an unsolicited Router Advertisement for each of R1’s routing prefixes
(unless R1’s prefixes are configured as “not advertised”).
• R2 responds to Neighbor Solicitations and (if configuration allows) Router Advertisements
on behalf of R1
• R2 terminates IPv6 traffic (such as pings) destined to R1’s SMLT IPv6 addresses
When R1’s RSMLT transitions back into the UP state and the HoldDown timer expires it
resumes IPv6 forwarding and R2 ceases to terminate IPv6 traffic on R1’s behalf.
Note that IPv6 allows a rich set of configuration options for advertising IPv6 routing prefixes
(equivalent to IPv4 subnets) and configuring hosts on a link. A prefix can be configured to be
or not to be advertised, to carry various flags or lifetime. These parameters affect how hosts
can (auto)configure their IPv6 addresses and select their default routers. Most relevant from
the RSMLT perspective is that an RSMLT node fully impersonates its peer’s IPv6 configuration
and behavior on the SMLT link – whatever its configuration happens to be. The above network
example illustrates one of the many possible deployment schemes for IPv6 routers and hosts
on a VLAN.
RSMLT provides both router failover and link failover. For example, if the Split MultiLink Trunk
link between R2 and R4 is broken, the traffic fails over to R1 as well.
Router R1 recovery
After R1 reboots after a failure, it becomes active as a VLAN bridge first. Packets destined
to R1 are switched, using the bridging forwarding table, to R2. R1 operates as a VLAN bridge
for a period defined by the hold-down timer.
After the hold-down time expires and the routing tables converge, R1 starts routing packets
for itself and also for R2. Therefore, it does not matter which of the two routers is used as the
next hop from R3 and R4 to reach IPv6 prefix 2003::/64.
When an IPV6 RSMLT peer recovers, the peer installs a temporary default route in the IPv6
routing table to point all the IPv6 traffic to the IST peer IP address for the hold down time.
(This is the same behavior as in IPv4 RSMLT.)
IPv6 RSMLT
Configuration — IPv6 Routing November 2010 71