> My idea with the draft is to address all the extensions (with all the 
> required technical detail) to enhance LISP+SDN deployments, that's the reason 
> for the name. Please keep in mind that this is just a -00 version and

I understand what you are trying to do but when you say "LISP+SDN" that can 
mean 100s of things. Here is an example:

(1) I have 1 ITR and 1000 ETRs.
(2) I have 100 groups where there membership is a subset of the 1000 ETRs.
(3) I don't want to do any protocol signalling to allow the ITR to replicate 
multicast packets for the 100 
groups to the various ETRs.
(4) I have an SDN application that programs the ITR with 100 map-cache entries.

So the above example can be classified as "LISP+SDN" but you, of course, had no 
intention to cover that in your draft. My point being is that you need to GET 
SPECIFIC.

> most of the content is still to be written. Despite the fact that the text 
> now refers mostly to flow aspects, my intention is to address other points 
> (such Publish/Subscribe mechanism, etc) on the near future.

Publish means that ETRs register 5-tuples to the mapping system. Subscribe 
means that ITRs send Map-Requests to find RLOC-records for those 5-tuples.

Unless you use specific language like I did above just saying "Pub/Sub" can 
mean dozens of things.

> I would prefer to keep all SDN related stuff here to avoid the proliferation 
> of drafts. If in the future we see

What is the "SDN related stuff" mean?

> that some parts (like what you call "LISP granular map-caches") require too 
> much technical detail, then we should consider the spin-off of those parts to 
> different drafts and keep this draft as a "SDN umbrella" (as Michiel 
> suggested).

If you take the details of the granular map-cache out of this spec, then what 
is the spec covering?

> At this point this draft is a tool to work on all the SDN specific aspects of 
> LISP and to generate some discussion.
> 
> Let's talk in London ;)

Where is the spec on SDN specific aspects of BGP and OSPF? There isn't any 
because it doesn't make sense.

Can someone else clarify this if you think I'm missing something?

Dino



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

Reply via email to