3Com S7906E Configuration Manual page 784

2) Upon receipt of this message, the peer is aware that the sending router is capable of Graceful
Restart, and sends an Open message with GR Capability to the GR Restarter to establish a GR
session. If neither party has the GR capability, the session established between them will not be
GR capable.
3) When an active/standby switchover occurs on a distributed device that acts as the GR Restarter,
sessions on it will go down Then, GR capable peers will mark all routes associated with the GR
Restarter as stale. However, during the configured GR Time, they still use these routes for packet
forwarding.
4) After the restart is completed, the GR Restarter will reestablish GR sessions with its peers and
send a new GR message notifying the completion of restart. Routing information is exchanged
between them for the GR Restarter to create a new routing table and forwarding table and have
stale routing information removed. Then the BGP routing convergence is complete.
MP-BGP
Overview
BGP-4 supports IPv4 unicasts, but does not support other network layer protocols like IPv6.
To support more network layer protocols, IETF extended BGP-4 by introducing Multiprotocol
Extensions for BGP-4 (MP-BGP) in RFC 4760.
Routers supporting MP-BGP can communicate with routers not supporting MP-BGP.
MP-BGP extended attributes
In BGP-4, the three types of attributes for IPv4 address format, namely NLRI, NEXT_HOP and
AGGREGATOR (AGGREGATOR contains the IP address of the speaker generating the summary route)
are all carried in updates.
To support multiple network layer protocols, BGP-4 puts information about network layer into NLRI and
NEXT_HOP. MP-BGP introduced two path attributes:
MP_REACH_NLRI: Multiprotocol Reachable NLRI, for advertising feasible routes and next hops
MP_UNREACH_NLRI: Multiprotocol Unreachable NLRI, for withdrawing unfeasible routes
The above two attributes are both optional non-transitive, so BGP speakers not supporting
multi-protocol ignore the two attributes and do not forward them to its peers.
Address family
MP-BGP uses address families to differentiate network layer protocols. For address family values, refer
to RFC 1700 (Assigned Numbers). Currently, the system supports multiple MP-BGP extensions,
including VPN extension and IPv6 extension. Different extensions are configured in respective address
family view.
1-15