IETF
taps
taps@jabber.ietf.org
Tuesday, November 11, 2014< ^ >
Room Configuration
Room Occupants

GMT+0
[22:56:32] Gorry Fairhurst joins the room
[22:57:46] csp joins the room
[22:58:47] MarieJose Montpetit joins the room
[23:00:14] Brian Trammell joins the room
[23:03:59] MarieJose Montpetit leaves the room
[23:04:39] MarieJose Montpetit joins the room
[23:06:07] Dave Thaler joins the room
[23:06:13] <Gorry Fairhurst> TSVWG now meet before and after lunch on THURSDAY :-)
[23:06:59] <Dave Thaler> I'll be temporary jabber scribe
[23:07:02] <Brian Trammell> *raises hand*
[23:07:02] Meetecho joins the room
[23:07:06] <Brian Trammell> hi dave! thanks
[23:07:15] <Dave Thaler> not sure who is the usual one
[23:07:26] <Dave Thaler> Agenda bashing
[23:07:54] hildjj joins the room
[23:08:24] <Brian Trammell> *waves*
[23:08:54] <Dave Thaler> TAPS in a Nutshell: Problem (slide 5)
[23:09:40] <Dave Thaler> slides: http://www.ietf.org/proceedings/91/slides/slides-91-taps-0.pdf
[23:09:52] iyengar@jabber.org joins the room
[23:10:10] <Dave Thaler> slide 6: TAPS Approach
[23:11:08] <Dave Thaler> slide 7: Deliverables
[23:12:17] <Dave Thaler> (handing scribing over to Jana)
[23:12:26] <Brian Trammell> hi Jana!
[23:12:43] <iyengar@jabber.org> Hello Brian!
[23:13:36] <iyengar@jabber.org> slide 8: Out of scope items
[23:14:28] <iyengar@jabber.org> slide 9: Milestones
[23:15:03] <Dave Thaler> Kevin Fall at mic
[23:15:11] <iyengar@jabber.org> (thanks, Dave)
[23:15:36] <Dave Thaler> terminology slides: http://www.ietf.org/proceedings/91/slides/slides-91-taps-1.pdf
[23:16:32] <iyengar@jabber.org> slide 2: Transport Service
[23:16:58] <iyengar@jabber.org> slide 3: Is a transport Service ...
[23:17:18] <iyengar@jabber.org> slide 4: Terminology
[23:17:49] <Dave Thaler> Mirja Külewind presenting
[23:17:57] <Dave Thaler> Kühlewind
[23:18:21] <Brian Trammell> and aaron falk on video :)
[23:19:03] resnick joins the room
[23:19:04] <iyengar@jabber.org> Is aaron on a video feed?
[23:19:18] <Brian Trammell> camera is pointed at the door behind aaron
[23:19:23] <iyengar@jabber.org> Dave Thaler at the mike
[23:19:23] <Brian Trammell> aaron's kinda in frame
[23:20:04] <Brian Trammell> mic: i'm okay with dave's suggestion.
[23:20:12] <Brian Trammell> define facility as facility
[23:20:48] <Brian Trammell> actually either is fine
[23:21:46] <iyengar@jabber.org> Brian — is video better?
[23:21:48] <Gorry Fairhurst> I think that Fcaility is used in at least svereal transport protocol  
[23:22:39] <Dave Thaler> Kevin Fall: see RFC 2126 for definitions
[23:23:16] <Brian Trammell> jana: yes
[23:23:25] <resnick> I propose "thingy".
[23:23:36] <iyengar@jabber.org> thingy is great
[23:23:53] PasiS joins the room
[23:24:09] <Brian Trammell> mic: transport service öppis
[23:24:16] <MarieJose Montpetit> we could use french too - truc, bidule, machin
[23:24:28] <Dave Thaler> the protocol component formerly known as Prince
[23:24:54] <Brian Trammell> (mic: more seriously i think we will be unable to avoid collisions)
[23:25:17] <Brian Trammell> actually scratch that we're moving on
[23:25:49] <iyengar@jabber.org> ok
[23:26:06] <Gorry Fairhurst> I agree - we can not avoid collisions - and I like mirja's terms
[23:26:13] cheshire joins the room
[23:26:31] <Brian Trammell> reliability without delivery is pointless
[23:26:54] <Brian Trammell> "yes, i applied the null operation to your message and it worked"
[23:27:18] <Brian Trammell> mic: what kevin is describing is what we called aspects
[23:27:27] <Brian Trammell> mic: aspect of feature would be fine
[23:27:51] <Brian Trammell> mic: "feature" is a thing that a protocol does on purpose, "aspect" is something that it does, whether on purpose or not
[23:28:01] <Gorry Fairhurst> Mic: Feature is OK, but that's another level of detail we need to do next
[23:28:41] <Brian Trammell> (thank you jana!!)
[23:29:04] <Dave Thaler> Ken Calvert at mic
[23:29:12] Glenn Scott joins the room
[23:29:18] <resnick> Brian, what's the diff between an aspect and a component (in the presented terminology)?
[23:30:01] <Dave Thaler> I'd like to hear an example of an "aspect" vs a "component" to better understand
[23:30:55] <iyengar@jabber.org> (happy to do it, Brian!)
[23:31:13] <Brian Trammell> pete: effectively none. we ditched "aspect" because it seemed too exotic
[23:31:29] <iyengar@jabber.org> Ah, so there's no aspect then
[23:31:42] <Brian Trammell> (in the article Joe and I wrote for IEEE Internet Computing, we used "dimension", but ditched _that_ because they're not orthogonal)
[23:32:12] <Brian Trammell> if we really want a noncolliding replacement for TS component, I would suggest aspect.
[23:32:14] <Dave Thaler> Kevin Fall from the floor: "characteristic" is the ISO 8072 term
[23:33:23] <Dave Thaler> Edward Lopez at mic
[23:34:01] <hildjj> The paper Brian is referring to: http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=6886121
[23:34:55] <Brian Trammell> i would also be okay with Andrew McGregor's "swap TS component and TP feature"
[23:35:22] <Brian Trammell> (I appear to be in an uncharacteristically agreeable mood today)
[23:35:35] <Gorry Fairhurst> I need to think on that ... Let's see if it works when we try to use it.
[23:35:58] <Dave Thaler> Ronald in 't Velt at mic
[23:36:45] <Brian Trammell> there is some other terminology IIRC that uses TS element
[23:36:51] <Dave Thaler> Jana Iyengar at mic
[23:39:27] <Dave Thaler> Pete Resnick at mic
[23:40:50] <Dave Thaler> Stuart Cheshire at mic
[23:43:14] <Brian Trammell> so this problem has always trying to match the crappy interface above the transport to the crappy interface below it
[23:43:30] <Brian Trammell> has always been
[23:44:03] <Brian Trammell> the terminology (and taps charter for that matter) assume the lower interface can be ignored
[23:44:36] <Dave Thaler> Andrew McGregor at mic
[23:44:40] <Brian Trammell> that might be impossible
[23:45:24] <iyengar@jabber.org> Nacho Solis at mic
[23:45:39] <cheshire> Ignacio Solis :-)
[23:46:38] <iyengar@jabber.org> Badge says Nacho Solis
[23:46:59] <Brian Trammell> so from a terminology standpoint i think we probably don't want to define a term for the controller...
[23:47:23] <Brian Trammell> but we might want to have a way to describe the things that the controller knows about the lower interface
[23:47:43] <Brian Trammell> transport aspects and path aspects maybe
[23:49:23] <iyengar@jabber.org> Hey Brian — do you want me to take that to the mic?
[23:49:30] <Brian Trammell> hm
[23:49:34] <Brian Trammell> yes please
[23:49:37] <iyengar@jabber.org> ok
[23:50:38] <iyengar@jabber.org> what "controller"?
[23:50:58] <Brian Trammell> the component in the middle nacho was talking about
[23:51:13] <Brian Trammell> or more generally "the thing that makes happy eyeballs etc work"
[23:51:46] <Dave Thaler> Kevin Fall at mic
[23:51:56] <Dave Thaler> (me at mic before Kevin)
[23:52:25] Meetecho leaves the room
[23:52:46] <Brian Trammell> or what Kevin is calling an agent
[23:53:55] <Brian Trammell> you're welcome, Mirja :)
[23:54:57] Meetecho joins the room
[23:55:51] <Dave Thaler> Toerless Eckert at mic
[23:57:02] <Dave Thaler> Andrew McGregor at mic
[23:58:20] <Dave Thaler> Michael Welzl at mic
[23:59:07] <Dave Thaler> Jana Iyengar at mic
[23:59:31] <Brian Trammell> or in terms of this terminology: the interface (to be designed in item 3) must allow the composition of protocols from (components/aspects) vs a transport protocol directly
[23:59:49] <Brian Trammell> i.e. i think the factoring of the terminology supports all of these things
Powered by ejabberd Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!