Hi Christian, Your message is actually spot on and very timely.
Last friday I have had a great meeting with Thomas and he actually suggested to decouple YANG from IMP transport. And I think this is great suggestion. In the same time main goal of IMP is to provide BGP-LS TLVs directly from IGP process over either TCP or QUIC sessions in native IGP format. So IMP is focusing on this very functionality. Many thx, R. On Sun, Jul 17, 2022 at 12:38 AM Christian Hopps <[email protected]> wrote: > > Robert Raszuk <[email protected]> writes: > > > Btw this independent attempt by two WG groups to normalize link state > > data is a clear proof that the YANG model has failed here. > > I'm not sure which "YANG model" you are referring to here (perhaps you > just mean YANG in general?), but I don't think YANG itself has failed at > all. Has anyone even tried to attack this problem directly using it yet? > > I'm pretty sure that GRPC is being used by google for telemetry, I don't > know if they're using it for LSDB information or not, but they have the > infrastructure setup for it obviously. I'm pretty sure Microsoft is doing > similar stuff to google with YANG modeled data as well. > > FWIW, YANG is a modeling language describing structured data, the > transport is just as important here, and NETCONF or RESTCONF probably > aren't the ones to use for your application. Maybe someone needs to look > into marrying the correct transport, with the correct YANG modeled data and > describe a system that would do what you want using existing technologies. > > Thanks, > Chris. >
_______________________________________________ Lsr mailing list [email protected] https://www.ietf.org/mailman/listinfo/lsr
