[12:13:13] --- Simon Leinen has joined
[12:18:45] --- Simon Leinen has left
[12:44:10] --- Simon Leinen has joined
[14:39:06] --- Simon Leinen has left
[15:08:05] --- Simon Leinen has joined
[15:52:43] --- stephen.morris has joined
[15:57:15] --- Simon Leinen has left
[15:58:35] --- dperkins3600 has joined
[16:02:06] --- badra has joined
[16:02:18] --- badra has left
[16:05:12] <dperkins3600> What is the confusion?
[16:06:43] --- ray has joined
[16:06:52] --- Okita has joined
[16:07:31] --- bert has joined
[16:07:34] --- atarashi has joined
[16:07:47] --- stephen.morris has left
[16:07:49] --- mg has joined
[16:07:50] <bert> Confusion?
[16:08:12] <bert> do you all hear the audio?
[16:08:13] --- lhotka has joined
[16:08:24] <dperkins3600> Audio is good today!
[16:08:29] <bert> Cause I was hoping os, so that I don't have to type too much
[16:08:41] <dperkins3600> No need to type for me
[16:12:53] --- mg has left
[16:13:28] <bert> agenda: http://www3.ietf.org/proceedings/07dec/agenda/netconf.txt
[16:15:15] --- mike has joined
[16:17:46] <bert> http://www3.ietf.org/proceedings/07dec/slides/netconf-3.ppt by Mohamad Badra
[16:18:20] --- Simon Leinen has joined
[16:18:45] <bert> slide 3
[16:19:27] <bert> slide 4
[16:20:19] <bert> be: access control is outside current charter
[16:21:02] <bert> dr: IESG wants early security review of this document, even before WG Last Call
[16:22:03] <bert> fair number of people have read the document (8-10)
[16:22:29] --- john.zhao has joined
[16:22:33] <bert> most of those think doc is ready to be accepted as WG document and work item
[16:22:44] <bert> nobody objects
[16:22:56] <bert> chair asks to pls review this document
[16:23:16] <bert> this needs to be taken to mlist as well
[16:23:50] --- badra has joined
[16:24:04] <bert> by Balazs
[16:24:17] <bert> http://www3.ietf.org/proceedings/07dec/slides/netconf-1.ppt
[16:24:29] <bert> slide 3
[16:24:38] <bert> slide 4
[16:25:15] <bert> slide 5
[16:25:53] <bert> slide 6
[16:26:08] <bert> slide 7
[16:26:37] <bert> slide 8
[16:26:50] <bert> slide 9
[16:27:12] <bert> slide 10
[16:29:10] <bert> slide 11
[16:29:34] <bert> slide 12
[16:30:46] <bert> slide 13
[16:31:04] <bert> sc: did send comments to mlist
[16:31:09] <bert> draft is right approach
[16:31:29] <bert> do we want to require XPATH for this capability?
[16:31:59] <bert> be: could be used from CLI, then we don;t care how they specify
[16:33:06] <bert> most implementations will support XPATH, so why would we make this optional
[16:33:12] <bert> Andy has a problem with that
[16:33:41] <bert> I am going to type less. Pls speak up if your audio is not good enough
[16:34:55] <bert> andy agrees with the 2 options
[16:35:09] <bert> dr: observation on use of YANG
[16:35:26] <dperkins3600> Audio is fantastic!
[16:35:41] <bert> we should probably not use this in the draft, since YANG is not ready yet
[16:36:16] <bert> phil:pls laborate on interactions between lock, get-config, commit etc
[16:39:06] <bert> this needs to be described more carefully
[16:39:44] <bert> wh: what if you change pieces of config outside your locked area?
[16:45:30] --- mg has joined
[16:45:41] <mg> Is there a way to list who has what locked?
[16:46:18] <mg> (I am not a netconf follower, more like auditing a class here, and don't want to embarass myself!)
[16:47:52] <bert> I believe that is part of the netconf monitoring effort
[16:47:59] <mg> Gotcha.
[16:48:02] <bert> But I can post the question here if you want
[16:48:12] <mg> I can too, just don't want to get in front of the mic :)
[16:48:27] <mg> I can research and send mail.
[16:48:32] <bert> ok
[16:48:56] <mg> Or not. Thanks, speaker :)
[16:49:03] <bert> so there we are. Sharon is working on monitoringdraft
[16:49:21] <bert> the speaker was sharon
[16:49:27] <mg> Thanks, Sharon.
[16:49:50] <bert> chairs asks: who read the drft
[16:50:13] <bert> quite a few read it
[16:50:21] <bert> 1 objection to adopt it
[16:51:06] <bert> http://www3.ietf.org/proceedings/07dec/slides/netconf-6.ppt by Mark Scott
[16:51:13] <bert> slide 2
[16:51:28] <bert> slide 3
[16:51:32] --- rohan has joined
[16:51:52] --- histerrier has joined
[16:53:31] --- histerrier has left
[16:53:37] <bert> slide 4
[16:54:29] <bert> slide 5
[16:55:17] <bert> slide 6
[16:56:47] <bert> slide 7
[16:57:12] --- mike has left
[16:57:37] <bert> slide 8
[16:58:48] <bert> slide 9
[17:00:15] --- stephen.morris has joined
[17:01:47] <bert> some 8 people read the doc
[17:02:07] <bert> 6 seem to think it is ready for adoption
[17:02:18] <bert> nobody objects
[17:03:01] <bert> http://www3.ietf.org/proceedings/07dec/slides/netconf-4.ppt
[17:03:12] --- stephen.morris has left
[17:03:16] <bert> slide 2
[17:03:30] --- rohan has left
[17:03:34] <bert> slide 3
[17:03:59] <bert> slide 4
[17:04:36] <bert> slide 5
[17:05:32] <bert> slide 6
[17:06:12] --- badra has left
[17:08:13] <bert> slide 7
[17:08:51] <bert> slide 8
[17:09:16] --- badra has joined
[17:10:32] <bert> slide 9
[17:11:04] <bert> slide 10
[17:12:46] <bert> slide 11
[17:13:15] --- badra has left
[17:13:50] <bert> slide 12
[17:15:03] <mg> As per typical, I've not read about this before today, but it seems that getting the WSDL from a specific device ties those RPC call prototypes to that device, and not to a more general network configuration model.
[17:17:30] <dperkins3600> Well, you really want to know with the device does and not what a standards body has published
[17:18:18] <mg> But this is a compile-time option for the developer -- if the device version changes, you have stale data. This is not a runtime dynamic thing, from my take on this.
[17:18:32] <dperkins3600> Oops - typo. That should have been "know what the device" and not "know with the device"
[17:19:39] <mg> It seems a method to advertise features in a monitoring tool discoverable way not compile time makes more sense.
[17:19:48] --- stephen.morris has joined
[17:19:51] <dperkins3600> Not saying "dynamic". Saying device specific and not industry general.
[17:20:53] <dperkins3600> Yes - Bert you said it correctly
[17:21:42] <bert> who read the document
[17:21:46] <bert> more than 10
[17:22:59] <bert> http://www3.ietf.org/proceedings/07dec/slides/netconf-5.ppt
[17:23:04] <bert> slide 2
[17:23:12] --- stephen.morris has left
[17:24:02] <bert> slide 3
[17:24:17] --- stephen.morris has joined
[17:25:02] <dperkins3600> Location is not operational workable if it is not the managed device. That is, IT MUST be the managed device. This is a problem with SNMP!
[17:26:05] <bert> slide 4
[17:26:11] <dperkins3600> Isn't this the same as the previous presentation?
[17:26:23] <bert> I believe it is presented as an alternative
[17:27:02] <bert> slide 5
[17:27:20] --- badra has joined
[17:27:37] <mg> previous was compile-time WSDL put-the-schema-in-your-binary, this one is ask-the-device-when-you-need-the-schema modle.
[17:27:38] --- badra has left
[17:28:27] <bert> slide 6
[17:28:39] <bert> dperkins: I do not understand your comment: Location is not operational workable if it is not the managed device. That is, IT MUST be the managed device. This is a problem with SNMP!
[17:28:56] <bert> should I mention it? And if so which slide (and talk) does it relate to?)
[17:29:16] <dperkins3600> Lets wait until questions.
[17:29:20] <bert> ok
[17:29:22] <mg> What if the device does not store the schema, this is a management issue as to where it is. And then you lose the connection between device and updated files on another (say) web server. Is this what you mean?
[17:29:23] <bert> slide 7
[17:30:58] <bert> slide 8
[17:31:00] <dperkins3600> Device manufacturers go out of business. It you return a URL to the out of business domain, you get nothing. Also, what the engineers do and what the "marketing people" do are not always in sync
[17:31:30] <mg> So storing the schema on the device is the right answer perhaps.
[17:32:11] <bert> slide 9
[17:32:18] <mg> You also don't want to depend on an external off-site link when you are trying to perhaps fix your routers.
[17:33:23] <dperkins3600> In SNMP you can return OIDs that are the AGENT capabilities. It has the problem of then figuring out how to map an AgentCap OID to a list of object and notification defs. Also, few agent developers saw any advantage of listing the capabilities. It was only more work.
[17:34:03] <dperkins3600> Yes - must store the device specific schema one the device.
[17:34:35] <dperkins3600> There must be an easy way to specify subset implementation.
[17:35:22] <dperkins3600> Note that this is being driven by management app developers.
[17:35:52] <dperkins3600> Don't believe that runtime update of management app is possible!
[17:36:18] <mg> I'm here to see how applicable netconf is to my universe. So far this fits more in line with how I would want schema discovery to work.
[17:36:42] <dperkins3600> Pls ask - how is subset schema implemented
[17:37:10] <mg> bert, are you willing to ask this for dperkins here?
[17:37:28] <bert> I can ask that, but there is a line, so be patient
[17:38:57] <dperkins3600> Ok - let me say it again - How do you specify that the device has implemented a subset of a "standard schema". Do you just return all of what is implemented, or is there a way to specify the delta?
[17:39:25] --- john.zhao has left: Computer went to sleep
[17:40:33] <bert> is it for this preso or previous
[17:44:10] <dperkins3600> Thanks Wes - well stated
[17:44:13] --- john.zhao has joined
[17:44:16] <bert> happy with the doiscussion?
[17:44:18] <bert> OK
[17:44:31] <dperkins3600> Thanks also Bert
[17:45:32] <dperkins3600> Pls - Must be on the device, since the location can go away, or out of sync of device and repository.
[17:46:10] <bert> I think several people have now stated that
[17:46:33] <bert> so Mark allows it to be on device but does not want to prescribe that (or so I understand)
[17:46:34] <dperkins3600> "Faith" has proven to be false hope!
[17:46:58] <mg> I can guess which device will be the more reliable :)
[17:47:08] <bert> this can be further discussed if it becomes aWG item
[17:47:15] <dperkins3600> The device that you are talking to!
[17:47:20] <mg> right
[17:47:21] <bert> some 10 or so people read this draft
[17:47:38] <bert> I think he does allow to store it on device that you are talking to
[17:48:05] <mg> It seems that it should be required to be on the device, but also perhaps available in an alternate, preferred location.
[17:48:32] <dperkins3600> I understand "allow" - I just believe the best (and only) choice is "must"
[17:48:41] <mg> Agreed.
[17:48:48] <bert> 5 want to accept Hidiki's draft asa WG item
[17:49:09] <bert> 7 want to adopt Marks draft as a WG work item
[17:49:20] <dperkins3600> How can this go forward without agreeing on a schema?
[17:52:42] <bert> think more about these 2 solutions and discuss on mlist
[17:54:39] <bert> http://www3.ietf.org/proceedings/07dec/slides/netconf-2.ppt
[17:55:37] --- mg has left
[17:57:40] --- lhotka has left
[17:58:55] --- Okita has left
[17:59:03] <dperkins3600> Does it implement any semantics?
[17:59:14] --- stephen.morris has left
[17:59:19] --- ray has left
[17:59:21] <bert> sorry dave session is closed
[18:00:05] <dperkins3600> Hey - would you ask David Partain to send me an email, so I can get his current email address. Mine is dperkins@dsperkins.com
[18:00:57] <bert> will do
[18:01:11] <dperkins3600> Thanks - signing off
[18:01:15] --- dperkins3600 has left
[18:02:23] --- john.zhao has left
[18:18:32] --- Simon Leinen has left
[18:18:37] --- atarashi has left
[18:21:29] --- bert has left
[19:05:21] --- Simon Leinen has joined
[19:56:17] --- Simon Leinen has left
[20:41:28] --- badra has joined
[20:44:00] --- badra has left
[20:49:05] --- Simon Leinen has joined
[22:34:08] --- Simon Leinen has left