Mailing List Archive

VPLS to multiple end points
Does anyone have any examples of doing a VPLS to multiple end points?

I have a use case where I want to bring in a couple of vlans to an ASR920
and then transport them via MPLS to several other ASR920s. So, basically a
layer-2 vpn.
Normally, I'd just bring in the vlans, and then do an xconnect on the
interface over to the destination. But, I'm struggling with how to send it
to multiple destinations.

Or, is there a specific reason to _not_ do this and do it some other way?

thanks

Shawn
_______________________________________________
cisco-nsp mailing list cisco-nsp@puck.nether.net
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/
Re: VPLS to multiple end points [ In reply to ]
Hey Shawn,

By definition all VPLS circuits are multipoint, VPWS (aka VLL's, xconnects)
is used for point-to-point designations.

Nuance aside, multipoint is a pretty common deployment strategy for L2VPN
services-- you'll want to look at using L2 VFIs for this.

A good reference point would be here:
https://www.cisco.com/c/en/us/td/docs/ios-xml/ios/mp_l2_vpns/configuration/xe-3s/mp-l2-vpns-xe-3s-book/mp-l2vpn-pseudo-swit-xe.html

Additionally, you can look at using BDI's if you need to provide routing
capabilities to the VPLS.

- Michael


On Thu, May 27, 2021 at 6:30 AM Shawn L <shawn@rmrf.us> wrote:

> Does anyone have any examples of doing a VPLS to multiple end points?
>
> I have a use case where I want to bring in a couple of vlans to an ASR920
> and then transport them via MPLS to several other ASR920s. So, basically a
> layer-2 vpn.
> Normally, I'd just bring in the vlans, and then do an xconnect on the
> interface over to the destination. But, I'm struggling with how to send it
> to multiple destinations.
>
> Or, is there a specific reason to _not_ do this and do it some other way?
>
> thanks
>
> Shawn
> _______________________________________________
> cisco-nsp mailing list cisco-nsp@puck.nether.net
> https://puck.nether.net/mailman/listinfo/cisco-nsp
> archive at http://puck.nether.net/pipermail/cisco-nsp/
>
_______________________________________________
cisco-nsp mailing list cisco-nsp@puck.nether.net
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/
Re: VPLS to multiple end points [ In reply to ]
Option 1 - VPLS can be manual (using only LDP and you *typing in the
neighbor ip addresses* to land the pw's on)

Option 2 - VPLS can be auto-neighbor discovered...
- BGP does the AD (auto discovery, that is, for the PE's participating in
the vpls)
- PW/LSP set up - can be signaled via BGP or LDP (RFC 4761 or 4762
respectively)

Of course with option 2 you need to fire up bgp on those pe's, I use a
private AS (64512-65535, pick one)... immediately consider RR, and dual RR
hub at that....

I thought I had some config from when I tested vpls on 920's a few years
ago, but I do have asr903 configs, which might be similar, just verify
before you go using these configs... sharing some stuff I had from my
previous lab tests...

This appears to be an RFC 4762 config...

lab-903-1 - 03.16.00.S

interface GigabitEthernet0/5/3
no ip address
negotiation auto
service instance 1 ethernet
encapsulation untagged

l2vpn vfi context v100
vpn id 10100
autodiscovery bgp signaling ldp
vpls-id 65535:10100
rd 10.10.12.246:32768
route-target export 65535:10100
route-target import 65535:10100
no auto-route-target

bridge-domain 100
member GigabitEthernet0/5/3 service-instance 1
member vfi v100

interface BDI100
ip address 1.1.1.6 255.255.255.0

router bgp 64512
bgp router-id 10.10.12.246
no bgp default ipv4-unicast
neighbor 10.10.0.254 remote-as 64512
neighbor 10.10.0.254 update-source Loopback0

address-family l2vpn vpls
neighbor 10.10.0.254 activate
neighbor 10.10.0.254 send-community extended
neighbor 10.10.0.254 prefix-length-size 2

NOTES.. Just like ME3600, ASR903 cannot change 4761 to 4762 within the vfi
on-the-fly.. Must delete and recreate.

eng-lab-903-1(config)#l2vpn vfi context v100
eng-lab-903-1(config-vfi)#autodiscovery bgp signaling ldp
% VFI auto-discovery signaling protocol changes are not allowed. Delete the
VFI and reconfigure.

eng-lab-903-1(config-vfi)#no l2vpn vfi context v100
% VFI removal failed. Please remove references to VFI before removal.

eng-lab-903-1(config)#do sh run | sec bridge-domain 100
bridge-domain 100
member GigabitEthernet0/5/3 service-instance 1
no member vfi v100


Also, again, I didn't like all PE's having same RD (auto) so I did
this.again, it worked without this, I mean it worked in auto RD mode, but I
did this anyway. Not sure what is best practice.

conf t
l2vpn vfi context v100
autodiscovery bgp signaling ldp
-> rd 10.10.12.246:32768



_______________________________________________
cisco-nsp mailing list cisco-nsp@puck.nether.net
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/