[c-nsp] When to switch to DFC3BXL

[Saku Ytti]

On (2007-05-18 00:36 +0200), Peter Salanki wrote:

> Take a look at the PPS in  "show platform hardware capacity" at peak  
> hours, when you start reaching 5-10mpps (depending on your  
> application) you should start thinking about adding DFCs. Also check  

5-10Mpps sounds really on the low side. It would be more than 10Mpps
even when doing MPLS and needing recirculate every 'packet'.
PFC w/o DFC and w/o recirculation is good for +30Mpps on IP path
and +20Mpps on MPLS path. This is what I got in lab, pushing
64bytes frames full duplex in 6704.

     Forwarding engine load:
                     Module       pps   peak-pps                     peak-time
                     5       29414904   29554159  13:09:14 EEST Thu Jun 8 2006
                     6       29415743   29444105  13:08:55 EEST Thu Jun 8 2006


> the bus utilization if you have non-fabric cards in the system.  
> Mixing fabric cards and non fabric cards is a really bad idea, as it  
> degrades the performance of the entire switch.
> 
> The biggest reason to add DFCs for most people is to get more space  
> for netflow entries.
> 
> Med vänliga hälsningar
> 
> Peter Salanki
> Teknikchef
> Bahnhof AB (AS8473)
> www.bahnhof.se
> Kontor: +46855577132
> Mobil: +46709174932
> 
> 
> 
> 17 maj 2007 kl. 03.55 skrev Chris Woodfield:
> 
> > "show platform hardware capacity"
> 
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.2.2 (Darwin)
> 
> iD8DBQFGTNjUiQKhdiFGiogRAgeyAJ93D+e0+XUlShzmwjLgmCHsidt9xgCfVjaK
> vPvYEC8JWCgrjR8QSlLGluA=
> =hK2c
> -----END PGP SIGNATURE-----
> _______________________________________________
> 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/

-- 
  ++ytti