04-02-2010 03:18 AM - edited 03-04-2019 08:00 AM
Bonjour,
In IP VPN, the VPN route label is an MPLS label, attached to a VRF.
In VPLS, the VPN route label is a PW label, attached to VSI.
Would you know why we can't use a PW label in an IP VPN?
Best regards,
Michelot
04-02-2010 05:04 AM
Hello Michelot,
actually the MPLS label space is only one for each node.
MPLS labels can be picked up by different signalling protocols from this single space and then advertised to peers in the appropriate sessions
in the case of EoMPLS pseudowire these labels are picked up and advertised by LDP over targeted LDP sessions
in the case of L3 MPLS VPN, MP BGP picks up one (or multiple ) labels per VRF site and will advertise them over VPNv4 address family to peers
All devices have to agree on the use of these service labels.
So when sending traffic to a specific PE node (external label pointing to device loopback) the internal service label is unique on that node and can be associated to an MPLS L3 VPN or to a pseudowire but not to both at the same time
Hope to help
Giuseppe
04-02-2010 06:21 AM
Hello Giuseppe,
Thanks for this reply.
> So when sending traffic to a specific PE node (external label pointing to device loopback)
> the internal service label is unique on that node and can be associated to an MPLS L3 VPN or
> to a pseudowire but not to both at the same time
Am I understanding correctly that?
For EoMPLS we have: ETH > PW (route label) > MPLS (tunnel label) > Transport layers
For IPoMPLS (RFC 4364) we have: IP > MPLS (route label) > MPLS (tunnel label) > Transport layers
So, my question was: why in IPoMPLS we don't have PW instead of MPLS for the route label?
Perhaps, the term MPLS for the VPN route label has evolved in PW after.
Thanks for an opinion.
Best regards,
Michelot
04-02-2010 06:52 AM
Hello Michelot,
real encapsulations on wire is:
>> For EoMPLS we have: ETH > PW (route label) > MPLS (tunnel label) > Transport layers
L2 Frame -> MPLS label (IGP,LDP BGP to next-hop)| service label= Pseudowire
I don't know why in many references the order of labels is reversed but the external label changes at each hope
L3 VPN
Eth --- MPLS label (IGP,LDP BGP to next-hop) | MP BGP VPN label
PW stands for pseudowire and was developed for EoMPLS scenario
Hope to help
Giuseppe
04-02-2010 08:00 AM
Hello Giuseppe,
Thanks for your words.
Subject (1)
>>> For EoMPLS we have: ETH > PW (route label) > MPLS (tunnel label) > Transport layers
> L2 Frame -> MPLS label (IGP,LDP BGP to next-hop)| service label= Pseudowire
> I don't know why in many references the order of labels is reversed but the external label changes at each hope
I agree that the external label is that one which change at each hope. So, I am used to write as above: (stable) PW over MPLS (modified hop to hop)
Perhaps, you are used to write: MPLS (modified hop-to-hop) over PW. Thus the client is below, crushed under all the layers. But, the result is same.
Subject (2)
> PW stands for pseudowire and was developed for EoMPLS scenario
It seems that a first RFC is RFC 3916 dated from september 2004.
PW was developped for ATM, FR, Ethernet: all L2, and also for CES.
Subject (3)
Perhaps, talking with you, I have an explanation for the difference between PW label and MPLS label for the route (the most iternal label).
RFC 3916 "These mechanisms (that is QoS mechanisms, DiffServ...) do not have to be defined in the PWE3 approaches themselves.
Best regards,
Michelot
04-02-2010 08:15 AM
Hello Michelot,
>> Perhaps, you are used to write: MPLS (modified hop-to-hop) over PW. Thus the client is below, crushed under all the layers. But, the result is same.
No, it is not the same.
I like to use the way the real encapsulation as it is on wire, and I personally don't like the other notation, there is a label stack and the inner label is that of the service and the external label in the encapsulation are usually created by different signalling protocols
If you make a packet capture you see the difference
A lot of presentations calls the external label in the encapsulation internal because it comes from an IGP
>> Perhaps, talking with you, I have an explanation for the difference between PW label and MPLS label for the route (the most iternal label).
I hope so
There are many RFCs on the subject.
I think everyone agrees on the fact that pseudowire is related to L2 transport services
Hope to help
Giuseppe
Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community: