Wireless Access

last person joined: 19 hours ago 

Access network design for branch, remote, outdoor, and campus locations with HPE Aruba Networking access points and mobility controllers.
Expand all | Collapse all

VRRP Scalability

This thread has been viewed 0 times
  • 1.  VRRP Scalability

    Posted Feb 05, 2012 09:01 PM

    I need to connect 4 active local M3 controllers to a backup controller all on the same switch.  I would like to use VRRP such that the backup controller is the standby unit for 4 VRRP instances.  I am doing this so I can have hitless failover when any one active controller fails.  Can an M3 controller scale to be the VRRP standby for 4 other controllers?  Can it maintain all the state information assuming the other 4 controllers are loaded to 75%?  Any insights would be appreciated.

     

    Thanks.



  • 2.  RE: VRRP Scalability

    EMPLOYEE
    Posted Feb 05, 2012 10:04 PM
    You would have better results making that controller the backup lms for all those other controllers. Vrrp is good but backup lms requires much less configuration and similar fail over.


  • 3.  RE: VRRP Scalability

    Posted Feb 05, 2012 10:31 PM

    In general, I agree.  However, in my case I need to provide hitless failover, that is affected users do not have to reassociate.   If I can provide a hitless failover using backup lms only, then I would do it. Is it possible?

     

    Thanks



  • 4.  RE: VRRP Scalability

    EMPLOYEE
    Posted Feb 06, 2012 04:26 AM

    Yes, as long as the backup controller puts the users on the same layer2 vlan you can accomplish this.

     



  • 5.  RE: VRRP Scalability

    Posted Feb 06, 2012 06:39 AM

    I can extend the VLANs to the backup controller and use the same VLAN assignment rules.  But  if I use backup lms, are the Pairwise Master Keys for user encryption and other required state information shared with the backup controller?  Would this be done via the Master Controller?

     

    Thanks.



  • 6.  RE: VRRP Scalability

    EMPLOYEE
    Posted Feb 06, 2012 06:46 AM

    For well-behaved applications like web browsing and email, this works very well.  I have never tested this with voice, but I don't expect it to work well  or even at all during a call.

     

    802.1x clients perform the best and they take between 7 and 30 seconds to fail over, depending.

     

    With regards to the PMK, it is not much different than having a client roam to another access point, so it works fairly well.  If you bridge clients to the same VLAN, the ARP table does not have to change, etc.

     

     



  • 7.  RE: VRRP Scalability

    Posted Feb 06, 2012 06:51 AM

    This is very good news for me.  Thank-you.