[j-nsp] EVPN all-active vs. layer 3

[adamv0025 at netconsultings.com]

> Rob Foehl
> Sent: Tuesday, November 12, 2019 3:19 AM
> 
> Given a pair of EX uplinked to a pair of MX, with various downstream CE
that
> may be single devices or their own layer 2 topologies, as in this terrible
> diagram:
> 
>    MX1 - MX2
>    |  / \  |
>    EX1   EX2
>       \ /
>       CEs
> 
> ...and a need to deliver various EVPN services to access ports on the EX,
> which currently run a bunch of layer 3 toward the CE devices (sometimes
> including VRRP across the CEs, or eBGP to them, or both):
> 
> Are there any "good" options for running all-active toward the EX while
also
> moving existing layer 3 up to the MX, or am I stuck with single-active and
> carving out IFLs for each use case?
> 
No you can never have active-active setup on L2 -because of the fundamental
limitation dictating that you can't have the one MAC address seen on two
ports at the same time.
In L3 I can have a two routes to a given prefix, each pointing via different
interface, you can't have the same with MAC addresses unfortunately.
This is due to conversational address learning and the absence of metric at
L2.

So that's why we have all these elaborate tricks to make it look as close to
active-active as possible while not breaking the fundamental rule.
Hence there are several levels of how close you want to get,  
ME to EX active-active I think it should be possible if you convert the bit
between MX-es and EX-es into L2 encapsulated in L3 or MPLS.
Say you run VXLAN between the MX-es and EX-es.
And then the bit between EX and CEs you can run MC-LAG (where one of the
ports in the LAG would be active and one will be backup). 

adam