[c-nsp] Solid L2 switch - 2948G or 3548-XL-EN?

[Julio Arruda]

Lamar Owen wrote:
> On Saturday 23 June 2007, Ian Dickinson wrote:
>> [The 2948G-L3] really was meant to be an L3 device - You had to do BVI's to
>> put a vlan on multiple ports.  The FE ports couldn't do ACLs - You had
>> a 'special command' to force traffic to route out of the GE port and back
>> again to apply ACLs there. 
> 
> Right.  It was somewhat like the 8500's in these regards, just less flexible.  
> So you do ACL's on the GE uplinks (something like you would have had to do 
> with the 8540, for instance, if the 16 port FE card didn't have the ACL 
> daughterboard) and don't do them on the FE ports.  For most usage that's not 
> to terrible of a limit.  The non-Catalyst-like VLAN setup is probably the 
> most odd thing about them (again, they are very much 8500-like in that 
> regard).
> 
> I was mostly curious if someone had had issues with them, other than these 
> documented limitations.


I assume the 2948G-L3 share the same heritage, the MMC EPIF forwarding 
brains ?
I was not even aware the 85xx had some 'sibling', but seems to be the case.
Does anyone has a table describing what switches use which NPUs or ASICs 
'off the shelf' or not ? I understand all 65xx PFC/DFCs is EARL (or the 
son/grandson of EARL), that was born in-house. Anything in the LAN side 
use any form of the PXF based gear ?