3com S7906E Manuel De Montage

 

1-15 

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. 

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. 

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: 

z 

MP_REACH_NLRI: Multiprotocol Reachable NLRI, for advertising feasible routes and next hops 

z 

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. 

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.