Re: [Isis-wg] 答复: Call for WG Adoption for draft-you-isis-flowspec-extensions-04

["Les Ginsberg (ginsberg)" <[e-mail]>]

> Again, this proposal just looks like configuring ones routers using a routing
> protocol. If that's what we are doing I'd like it to be more clearly understood,
> and then we can decide if this is what we think routing protocols should be
> doing. :)
>

I do NOT think this is what routing protocols should be doing. Our job is to send information needed for routing. 

If you want to use the IS-IS flooding mechanism as a general purpose transport then you should use GENAPP (RFC 6823) in combination with MI (RFC 6822). IT is then clear that you are defining application information and this falls under TLV 251 using an IANA assigned application identifier.

Alternatively you could use other mechanisms (such as Chris suggests below) - which is also fine and likely a better choice in this case. But this is NOT something that belongs in an IGP instance used for routing - and certainly not in router capabilities.

   Les


> -----Original Message-----
> From: Isis-wg [mailto:[e-mail]] On Behalf Of
> [e-mail]
> Sent: Friday, March 04, 2016 2:54 PM
> To: Youjianjie
> Cc: Marc Binderberger; [e-mail]
> Subject: Re: [Isis-wg] 答复: Call for WG Adoption for draft-you-isis-flowspec-
> extensions-04
> 
> 
> Youjianjie <[e-mail]> writes:
> 
> > Hi Chris,
> >> I agree if moving forward with this I think we need to look at GENAPP as
> well..
> >>
> >> Firs though I'd like to hear more on why intra-domain flowspec (or
> >> the
> >> equivalent) can't simply be defined using a yang model.
> >
> > Do you mean disseminating FlowSpec rules by using I2RS? I think it could be
> an alternative. But They each have their pros and cons. If using routing
> protocol to disseminate FlowSpec rules, when networks change, FlowSpec
> rules could be flexible as route entries, i.e. without involvement of
> management devices.
> 
> 
> I mean one would use netconf (or restconf, ...) and flowspec yang model to
> configure ones routers with the flowspec data.
> 
> Can you explain more what you mean when talking about network changes?
> The flowspec entries aren't changing when network changes occur.
> If the originator changes them it knows it needs to configure the new entries
> on the routers.
> 
> Again, this proposal just looks like configuring ones routers using a routing
> protocol. If that's what we are doing I'd like it to be more clearly understood,
> and then we can decide if this is what we think routing protocols should be
> doing. :)
> 
> Thanks,
> Chris.
> 
> >
> > Thanks,
> > Jianjie
> >
> >> Thanks,
> >> Chris.