3Com Switch 8800 Advanced Software V5 Configuration Guide

478 CHAPTER 36: IPV6 BGP CONFIGURATION
Tuning and
Optimizing IPv6 BGP
Networks
This section describes configurations of IPv6 BGP timers, IPv6 BGP connection soft
reset and the maximum number of load-balanced routes.
■ IPv6 BGP timers
After establishing an IPv6 BGP connection, two routers send keepalive messages
periodically to each other to keep the connection. If a router receives no keepalive
message from the peer after the holdtime elapses, it tears down the connection.
When establishing an IPv6 BGP connection, the two parties compare their
holdtimes, taking the shorter one as the common holdtime. If the holdtime is 0,
neither keepalive massage is sent, nor holdtime is checked.
■ IPv6 BGP connection soft reset
After modifying a route selection policy, you have to reset IPv6 BGP connections to
make the new one take effect, causing a short time disconnection. The current
IPv6 BGP implementation supports the route-refresh feature that enables dynamic
IPv6 BGP routing table refresh without needing to disconnect IPv6 BGP links.
With this feature enabled on all IPv6 BGP routers in a network, when a routing
policy modified on a router, the router advertises a route-refresh message to its
peers, which then send their routing information back to the router. Therefore, the
local router can perform dynamic routing information update and apply the new
policy without tearing down connections.
If a router not supporting route-refresh exists in the network, you need to
configure the peer keep-all-routes command on the router to save all route
updates, and then use the refresh bgp ipv6 command to soft-reset IPv6 BGP
connections.
To do... Use the command... Remarks
Enter system view system-view -
Enter BGP view bgp as-number Required
Enter IPv6 address family view ipv6-family -
Allow the local AS number to
appear in AS_PATH of routes
from a peer/peer group and
specify the repeat times
peer { ipv6-group-name |
ipv6-address } allow-as-loop
[ number ]
Optional
Not allowed by default
Specify a fake AS number for
a peer/peer group
peer { ipv6-group-name |
ipv6-address } fake-as
as-nmber
Optional
Not specified by default
Neglect the AS_PATH
attribute for best route
selection
bestroute as-path-neglect Optional
Not neglected by default
Configure to carry only the
public AS number in updates
sent to a peer/peer group
peer { ipv6-group-name |
ipv6-address } public-as-only
Optional
By default, BGP updates carry
private AS number
Substitute local AS number
for the AS number of a
peer/peer group indicated in
the AS_PATH attribute
peer { ipv6-group-name |
ipv6-address } substitute-as
Optional
Not substituted by default