[c-nsp] BRAS Redundancy

Arie Vayner (avayner) avayner at cisco.com
Mon Feb 22 14:52:26 EST 2010


Usually for static IP assignments you would have to redistribute the
connected/static (static for routes) prefixes into the routing protocol
(I would recommend BGP) so that you advertise them as /32. No magic...


-----Original Message-----
From: cisco-nsp-bounces at puck.nether.net
[mailto:cisco-nsp-bounces at puck.nether.net] On Behalf Of Anthony McGarry
Sent: Monday, February 22, 2010 13:50
To: cisco-nsp at puck.nether.net
Subject: [c-nsp] BRAS Redundancy


I was hoping someone can help me with the following issue.

I currently have a 7301 acting as my BRAS running on 12.2(33)SRD3. I use

the ISG feature to terminate PPPoE sessions on QinQ subinterfaces.

The virtual templates associated with the bba groups use ip unnumbered 
loopback 0.
The IP on loopback 0 is x.x.96.1/21

DHCP is configured for client IP address assignment using DHCP pools as 
relay agents to a central DHCP server.

ip dhcp pool DHCP
   relay source x.x.96.0
   class DHCP
      relay target x.x.111.5

I would now like to install a second 7301 for load balancing/redundancy.

I currently trunk the QinQ vlans to the existing 7301 so I just do the 
same for the second 7301.

On the second 7301 I assign a new /21 network for DHCP assignment.
This works fine for dynamic IP assignment.

My problem is that we have multiple customers with static IP address 
assignment from the DHCP server.

How can I assign the same IP address to a certain client session if they

login to either BRAS when each BRAS has a unique network associated with

the loopback 0 interface.

I was thinking mobile IP but I have not tested in the lab and not sure 
if it is a supported solution.


cisco-nsp mailing list  cisco-nsp at puck.nether.net
archive at http://puck.nether.net/pipermail/cisco-nsp/

More information about the cisco-nsp mailing list