Re: [irs-discuss] Thoughts on draft-ward-irs-framework

K - so the idea of the IRS sub-interface to the RIB being semantically
similar to another routing process sounds like an initially useful way
of describing it - but we're going to want to qualify out what that
means in terms of interactions and into the appropriate description &
models.

Alia

On Thu, Aug 2, 2012 at 1:26 AM, Russ White <russw@riw.us> wrote:
>
>> ----> [Russ]: +1 - This Offboard controller processing aligns with the ONF view of SDN.  And it has been done in practice for 10+ years. However as you know, the timing between the calculation and download to HW is important.
>
> If we just insert the information into "yet another process" parallel
> with the already existing routing processes on the hardware, I think we
> shunt aside the problems with insertion speed, etc. Routes installed in
> this "off board interface routing process" would be handled just like
> routes learned through other sources. Any improvements in the speed of
> route insertion from, say, OSPF running on the box would be reflected in
> speed improvements for IRS. IMHO, this is a huge win overall.
>
>> One
>> difference, though, is the idea that different static routes installed
>> by IRS may want to have different preference values.
>
> All the implementations I've worked with have the ability to install
> routes from a single process with different preferences. If no
> preference is specifically stated, the process' preference is used.
> Specific per route preferences override these, though --see, for
> instance, the EIGRP distance command in IOS. I suspect all
> implementations have the ability to handle this sort of thing in the API
> between each routing process and the RIB, but it's just not used a lot.
>
> :-)
>
> Russ
> _______________________________________________
> irs-discuss mailing list
> irs-discuss@ietf.org
> https://www.ietf.org/mailman/listinfo/irs-discuss