Greetings, all,
We've integrated the FLUTE/ALC text into the working copy of the document
(github.com/britram/taps-transports) as well as moving over to a slightly more
streamlined Makefile (though, due to smart quote replacement issues I'm still
debugging -- don't ask -- the toolchain to build the RFC from the markdown is
not working at the moment).
We've discussed the decisions taken in the Prague meeting about how to handle
features vs. components in each transport protocol section. Our (somewhat
complicated) attempt on -06 to really describe TCP in terms of what we believe
the components are didn't work out so well, and we'll go back to the previous,
simpler, more thoroughly decomposed list from the previous version.
It strikes us that part of the issue is that what we've listed for each
protocol in the 3.x.3 sections aren't really components (in the sense of sets
intertwined of behaviors that are inherent to the protocol as designed), but
rather features (whether or not these features should be exposed in the
eventual TAPS framework). The *distinction* between a component and a feature,
though useful for thinking about how to decompose protocols, is (as Michael
indicated in Prague) perhaps a bit academic when the point of the exercise in
the first place is to ensure we have a reasonably complete list of features
that TAPS can/should support.
So we propose the following resolution to the components-versus-features
confusion on a per-protocol basis:
(1) Keep each list in the 3.x.3 sections more or less as-is.
(2) Retitle the sections 3.x.3 "Features Provided" and make editorial changes
as necessary for agreement.
Please yell if you think this is a terrible idea; otherwise we'll spin a -07
shortly. After this, we still need to handle the following items
• Karen & Michael will add some text on SCTP
This is actually Mirja's action; she'll review the SCTP section and coordinate
with Karen & Michael shortly. Arms untwisted 'til then :) and apologies -- I
thought we were synced up here.
• Varun will add some text on RTP
We're also missing text on WebSockets. Since we have nothing for either at this
point, we'll strike these sections
• Gorry will write something on applications learning about network
conditions (ICMP)
This is still pending.
We'll integrate these into an -08, and then I think we're ready for WGLC.
Thanks, cheers,
Brian and Mirja
_______________________________________________
Taps mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/taps
