[nsp] ATM design issues (PVP/PVC redundancy/failover)

[Michel Renfer]

Hi Paul

ATM itself cannot handle fail-over in your case. ATM VCs (PVC, SVCs
and Soft-PVCs) are always configured from NSAP (ATM) address A to
NSAP address B. 

So if you need failover, you need to cover that on an upper layer. 
Create VCs to both 7206 and solve then the failover with a routing 
protocol. Or if you use the VCs for aggregating L2TP tunnels (DSL users)
you need to talk with the telco about session balancing.

cheers,
michel

> -----Original Message-----
> From: Paul Grehan [mailto:paulg at nt.com.au] 
> Sent: Thursday, May 13, 2004 9:14 AM
> To: cisco-nsp at puck.nether.net
> Subject: [nsp] ATM design issues (PVP/PVC redundancy/failover)
> 
> Hi,
> 
> I'm looking into having a failover solution in place for a 
> hardware failure
> in the following scenario and was wondering if anyone had an 
> opinion on how
> best to address it.          
> 
> I have PVP's (six) arriving on a single access that I need to 
> terminate over
> two 7206's
> 
> These terminations are distributed according to load, but I 
> need to have a
> recovery solution in place in the event of a hardware failure 
> on either of
> these 7206's.
> 
> The likelihood may be limited, but I'd like to see if I could 
> use some sort
> of PNNI soft PVC functionality, or if only a UNI solution is 
> viable, a point
> to multipoint solution that would allow a simple change over 
> in event of
> failure... 
> 
> 6 x PVP's           3 x PVP's to 
>                     each 7206
> 	                              |-----------|
> 	  --------                    |   7206    |
> 	 |        | ----------------->|-----------|
> -----> | LS1010 |                    
> 	 |        | ----------------->|-----------|
> 	  --------                    |   7206    |
> 	                              |-----------|
> 
> I will happily split the PVP's into corresponding PVC's in 
> the switch if
> necessary...
> 
> Sorry if this is a very fundamental thing - I am also aware 
> of the other
> failure point being the LS1010, which I think would have much 
> less chance of
> failure...
> 
> TIA
> 
> Paul Grehan
> 
> _______________________________________________
> cisco-nsp mailing list  cisco-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/cisco-nsp
> archive at http://puck.nether.net/pipermail/cisco-nsp/
> 
>