[c-nsp] ASR vs 6807

[Chris Welti]

Hi Simon,

you can also do port-to-subint on the Sup720 using ethernet interworking:

one end:

interface TenGigabitEthernet3/2
  xconnect y.y.y.y 1 encapsulation mpls
end

the other:

interface TenGigabitEthernet4/2.2010
  encapsulation dot1Q 2010
  xconnect x.x.x.x 1 pw-class atom-eth-iw
end

The magic is in the pw-class atom-eth-iw:

pseudowire-class atom-eth-iw
  encapsulation mpls
  interworking ethernet

We're using that in production and it works just fine.

Cheers,
Chris

On 27/11/14 12:10, Simon Lockhart wrote:
> In simple terms (and I apologise if this is fixed in Sup2T, as most of my
> experience has been on the Sup720), with the 6500/6800 platform, you can only
> do port-to-port or subint-to-subint VPWS, but not port-to-subint (which you can
> on the more capable boxes, or with the ES cards on the 6500/6800).
>
> Simon
>
> On Thu Nov 27, 2014 at 11:05:18AM +0000, R LAS wrote:
>> Hi Simon
>> can you detail more "ASR9k can be more flexible on EoMPLS (VPLS) than 6807" ?
>>
>> Regards
>>
>>> Date: Thu, 27 Nov 2014 10:26:55 +0000
>>> From: simon at slimey.org
>>> To: dim0sal at hotmail.com
>>> CC: cisco-nsp at puck.nether.net
>>> Subject: Re: [c-nsp] ASR vs 6807
>>>
>>> On Thu Nov 27, 2014 at 10:18:41AM +0000, R LAS wrote:
>>>> Discussing a new architecture of DCI (Data Center Interconnection), Cisco
>>>> raccomends both ASR9k and 6807.  The architecture requested by the customer
>>>> forecast MPLS/VPLS supported by DCI.
>>>>
>>>>  From pricing point of view there is a quite big difference (win 6807), from
>>>> feature point of view Cisco says the difference is "only" the number of
>>>> mac-addresses supported and the sw modularity.
>>>>
>>>> Can anybody help in digging more the "technical" difference ?
>>>
>>> I'm going through much the same at the moment, and settling on 6807, largely
>>> from a price perspective.
>>>
>>> ASR9k is (today) a more capable box for routing - particularly if you want
>>> higher bandwidths. ASR9k has 100G ports today. 6807 only has 40G. ASR9k can
>>> be more flexible on EoMPLS (VPLS) than 6807.
>>>
>>> 6807 has a lot of potential (880G per slot), but it's not supported by either
>>> Supervisors or Linecards that are available today (current limit is 80G/slot).
>>>
>>> Simon
>>