Re: [Detnet] A question about draft-gandhi-mpls-stamp-pw

Hi Greg, Stewart,

There are two use-cases that can benefit with STAMP packet without using
IP/UDP header and using STAMP GACH type:

*Use Case 1: *
•Same MPLS Encapsulation as Data Traffic carrying Control-Word (CW) for
monitoring ECMP and fate sharing
1.MPLS Ethernet PW [RFC4448]
2.L2-Specific Sublayer (L2SS) used in L2TPv3 PW carry CW (no IP header)
[RFC3931]
3.PLE [draft-ietf-pals-ple] traffic over PSN (as VPWS) carry CW (no IP
header)
4.TDM over IP [RFC5087] payload (structure-aware/structure-agnostic) carry
CW with or without IP header

*Use Case 2: *
•Demultiplex Control Channel (VCCV/OAM)
1.G-ACh types allow to demultiplex *VCCV Control Channel *for PWs [RFC7708]
and MPLS-TP PWs [RFC7189]
•For example, G-ACh types for BFD packets with or without IP/UDP headers
allow to demultiplex VCCV Control Channel [RFC5885]
2.G-ACh types allow to demultiplex *OAM Control Channel *for MPLS-TP LSPs
and PWs [RFC5960]

Hope this clarifies.

Thanks,
Rakesh


On Mon, Nov 8, 2021 at 4:05 AM Stewart Bryant <[email protected]>
wrote:

> I am inclined to wonder the same.
>
> If it is about not having an IP address to hand, I seem to recall that one
> mode of use over 0x21 was to use local host.
>
> Stewart
>
> Sent from my iPad
>
> On 7 Nov 2021, at 21:01, Greg Mirsky <[email protected]> wrote:
>
> 
> Dear Authors,
> I've read the draft and would appreciate your clarification to my question:
>
>    - What benefits do you see in defining two new ACH types for STAMP
>    protocol over a PW?
>
> I believe that there are no apparent advantages of using PW, i.e., without
> IP/UDP headers, encapsulation compared to using IP/UDP. The latter can be
> easily done using ACH types 0x0021 and 0x0057 for IPv4 and IPv6 address
> family respectively.
>
> What do you think?
>
> Regards,
> Greg
> _______________________________________________
> detnet mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/detnet
>
> _______________________________________________
> Pals mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/pals
>

_______________________________________________
Pals mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/pals