Re: [irs-discuss] IRS Problem Statement Posted

[Alia Atlas <[e-mail]>]

Yes, I've heard the comments that ForCES may have good characteristics
for the protocol aspects of IRS.  I think we need to focus on
discussing the problem and use-cases - before we start seriously
considering protocols.

More information about them and their potential relevance is always welcome.

Alia

On Tue, Jul 31, 2012 at 12:40 PM, Jamal Hadi Salim <[e-mail]> wrote:
> On Tue, Jul 31, 2012 at 12:30 PM, Alia Atlas <[e-mail]> wrote:
>
>> This is one of the key differences - ForCES is about programming the
>> forwarding plane.  IRS is about interfaces to the routing system - the
>> lowest it goes is to the RIB layer.  IMHO, this gives good control and
>> abstraction - no need to model forwarding plane differences.  That's
>> what the router OS is for.
>
> Ok, guess i am going to have say something then;->
> While ForCES did start with intent to do control-datapath interfacing,
> given that:  we have a data model (and language); a protocol that is
> agnostic of what the entity being configured is - over the last few
> years it has been used for anything that can be modelled using
> the ForCES language. Basically model it as an LFB and you are good
> to go.
> I am actually presenting on how we do parts of the FE management
> plane via ForCES at a session today. In our implementation, this
> is how we do remote debug settings, change syslog levels etc.
> Nothing to do with a FIB table programming etc...
>
> cheers,
> jamal