Re: [ippm] AD Review: draft-ietf-ippm-twamp-session-cntrl-03

[Lars Eggert <[e-mail]>]

Hi,

On 2010-2-26, at 8:58, Al Morton wrote:
> When we went ahead with the Mixed Security Mode RFC for TWAMP,
> http://tools.ietf.org/html/rfc5618
> we made it an Update for RFC 5357.  
> 
> But it's not *mandatory to implement* RFC 5618 if you are building
> RFC 5357 TWAMP, AFAIknew.  Is that what we've effectively done
> by making it an Update?

what the "Updates" header means is unfortunately not well defined. To some, "Updates X" in document Y means "if you have implemented X, you now also need to implement Y to remain compliant." To others, it just means "if you implemented X you probably now also want to read Y".

The IESG is working on a statement that hopes to clarify the different interpretations. Personally, I'm more of a subscriber to the first viewpoint.

For RFC5618, we have no issue, because although it says that it updates RFC5357, it's very clear from the text that it is an OPTIONAL extension. In other words, it's more or a "also read this" update.

> We want to make sure that someone who builds TWAMP knows that there
> have been new (optional) features added, like mixed security mode and 
> Individual Session control, etc. Update seemed like the right mechanism.

We can do the same here, also for consistency: add an "Updates: 5357" header, but then in the text make it very clear that this is an OPTIONAL extension that is not required to implement for TWAMP compliance.

Lars