IETF
dhc
dhc@jabber.ietf.org
Tuesday, November 5, 2013< ^ >
tomek has set the subject to: Next DHC meeting: IETF'87 Berlin
Room Configuration
Room Occupants

GMT+0
[00:13:35] Bernard Volz joins the room
[00:22:40] Bernard Volz leaves the room
[15:33:40] Bernard Volz joins the room
[15:33:56] Bernard Volz leaves the room
[16:17:43] Ralf Weber joins the room
[16:18:01] Ralf Weber leaves the room
[16:18:18] Ralf Weber joins the room
[16:40:36] Bernard Volz joins the room
[16:40:49] Ralf Weber leaves the room
[16:41:32] <Bernard Volz> Hi … welcome to IETF-88 & DHC WG meeting
[16:48:41] <Bernard Volz> we will start in about 12 minutes.
[16:54:30] Dan York joins the room
[16:54:41] Dan York has set the subject to: Next DHC meeting: IETF'88 Berlin
[16:55:44] <Bernard Volz> We are in vancouver.
[16:57:10] danny joins the room
[16:57:43] <danny> is audio working?
[16:57:45] <Bernard Volz> if anyone is in the dhc wg room and willing to be a jabber scribe, please let me know.
[16:58:05] <Bernard Volz> did you hear the audio test?
[16:58:26] <Bernard Volz> we will start in a few minutes.
[16:58:46] <danny> no
[16:59:20] Ole Troan joins the room
[16:59:45] <Bernard Volz> did you hear that?
[16:59:48] <danny> I don't think audio is working.
[17:00:38] pebersman joins the room
[17:00:56] <danny> I just tried a different room and got nothing
[17:01:02] <Ole Troan> now I can hear
[17:02:06] Ted Lemon joins the room
[17:04:44] Jinmei Tatuya joins the room
[17:05:37] <Bernard Volz> is audio ok now?
[17:06:07] Suzanne Woolf joins the room
[17:07:54] Ian Farrer joins the room
[17:08:48] <Ole Troan> yes, audio is OK
[17:09:15] <danny> hmm, it's strange that I am receiving nothing at all
[17:09:28] <Ole Troan> Danny, you are on channel 4?
[17:09:44] <danny> yes.
[17:09:55] <Ted Lemon> Have you turned it off and then on again?
[17:10:03] <danny> I wonder if our firewall is blocking it?
[17:10:11] <danny> yep
[17:12:18] <danny> I restarted UMPlayer and redid the link so it may be the corporate firewall
[17:15:11] <Bernard Volz> tomek is presenting now on 3315bis work.
[17:24:18] Brian Haberman joins the room
[17:27:29] Ralf Weber joins the room
[17:29:01] <Ted Lemon> Ole, where is your remote robot camera today?
[17:29:04] <Bernard Volz> now on to dhcpv4-over-dhcpv6 (qi sun presenting).
[17:29:38] <Bernard Volz> join in appear.in/dhc?
[17:31:29] <Ole Troan> does that scale to many I wonder?
[17:31:39] <Ted Lemon> Bernie, I'm not sure that's a good use of your laptop, but it's certainly cool.
[17:33:38] <Ole Troan> how could he possibly better use his laptop. ;-)
[17:34:22] sarikaya2012 joins the room
[17:34:25] <Ted Lemon> Taking notes on what to do next?   Dunno.
[17:35:39] Myung Hee Choi joins the room
[17:36:48] Andrew Yourtchenko joins the room
[17:37:10] tomtaylor joins the room
[17:37:41] <Ole Troan> wonder why this isn't a dead lock
[17:38:11] <tomtaylor> Current topic?
[17:38:20] <Ole Troan> yes
[17:38:41] <tomtaylor> Sorry, what is the curerent topic?
[17:38:41] <Ole Troan> tightly coupling DHCPv4 and DHCPv6 I mean
[17:38:42] <tomtaylor> ah
[17:38:44] ogud joins the room
[17:43:22] <Ted Lemon> It's really no different than something like PPCP/IPCP.
[17:43:56] <Ole Troan> Ted: how come? there is no interdependencies there?
[17:44:29] <Ted Lemon> I thought you had to do PPCP before you could do IPCP.   Maybe I'm misremembering the acronyms.
[17:44:29] <Ole Troan> well, I guess we can do it on the mailing list.
[17:46:11] <Ole Troan> So can this be used to provision a client on Ethernet then?
[17:46:46] <Ted Lemon> The protocol is running native over the ethernet, but it's providing configuration information for a tunnel.
[17:47:24] <Ole Troan> No, since this isn't restricted to any specific link-layer, can I use e.g. DHCPv4 over DHCPv6 to configure a client connected to an Ethernet?
[17:47:45] mcharlesr joins the room
[17:47:54] <Ted Lemon> Yes, if you can figure out a context in which that makes sense!
[17:48:00] <Ted Lemon> That's the point.
[17:48:07] <Ole Troan> Little of this makes sense, which is my point.
[17:48:28] <Ole Troan> Which is why I make the argument that you run DHCPv4 over the link-layer (aka tunnel) pretty much unchanged.
[17:48:32] <Ted Lemon> What is the outcome that you are advocating, Ole?
[17:49:02] <Ole Troan> DHCPv4 over the link-layer, DHCPv4 unchanged. No new protocol  or options.
[17:49:10] <Ole Troan> DHCPV6 to provision tunnel
[17:49:27] <Ole Troan> link-layer == tunnel
[17:49:35] <Ted Lemon> DHCPv4 over a port-restricted tunnel is _not_ unchanged, and only works if the tunnel works prior to IPv4 address configuration.
[17:49:37] <Ole Troan> Run DHCPv4 just like you do on any other link-layer.
[17:49:54] <Ole Troan> With RFC5107 it doesn't have to be treated as port restricted
[17:50:03] <Ole Troan> The tunnel is provisioned with DHCPv6
[17:51:06] <Ole Troan> The only thing that needs changing is the relay, which needs to support a new option 82 format.
[17:51:28] <Ole Troan> No new DHC documents required, possibly only a "this is how you put things together" document
[17:51:29] Ian Farrer leaves the room
[17:52:18] <Ole Troan> Ted: This is IPv4 we are talking about. What could possibly be the other "customers" for this?
[17:52:36] <Ole Troan> And all of these are mechanisms to sunset IPv4.
[17:54:57] <Ted Lemon> Ole, that's simply not a valid question.   We can't enumerate all things that may be discussed in the future.
[17:55:14] <Ole Troan> Exactly, so I don't understand why you bring up imaginary link-layers? ;-)
[17:55:41] <Ted Lemon> Because this has to account for them.
[17:55:46] <Ole Troan> Why?
[17:55:55] <Ole Troan> Can we even imagine a link-layer here?
[17:56:03] <Ted Lemon> Because it's a solution that we recommend for all future applications of DHCPv4 in an IPv6 island.
[17:56:03] <Ole Troan> Or get a requirement from softwire for one?
[17:56:04] <Ted Lemon> So it has to work.
[17:56:13] <Ted Lemon> This is not about softwire.
[17:56:30] <Ole Troan> Where does this requirement come from them?
[17:56:38] <Ole Troan> And what is it exactly?
[17:56:50] <Ted Lemon> It's come up several times in different contexts in the DHC working group over the past six years or so.
[17:56:56] <Ted Lemon> It will likely come up again.
[17:57:00] <Ole Troan> That you may in the future have a link-layer that does not support transport of DHCPv4, but does support transport for IPv4?
[17:57:16] <Ole Troan> Can you cite some references for that? Cause I'm a bit lost there.
[17:59:04] Ian Farrer joins the room
[17:59:11] <Ted Lemon> No, there have been requests to add IPv4 support to DHCPv6.
[17:59:22] <Ted Lemon> Because people wanted some IPv4 configuration in an IPv6 island.
[17:59:23] <Ole Troan> Correct. And that applies to this problem how?
[17:59:39] <Ole Troan> Do we agree that IPv4 configuration without IPv4 transport is meaningless?
[18:00:01] <Ted Lemon> The requirement is for a mechanism for doing IPv4 configuration in an IPv6 island.   Not IPv4 configuration of softwire tunnels.   That's an example use case.
[18:00:26] <Ole Troan> Where is that requirement from? What problem are you trying to solve? Do you agree with the statement above?
[18:00:40] <Ted Lemon> I am not even _interested_ in the statement above, Ole.
[18:00:48] 934ea03a4ea9ce8c joins the room
[18:00:59] <Ole Troan> Interesting. This really seems like basis for an appeal.
[18:01:09] <Ted Lemon> I am interested in closing out this issue.   If you have some technical objection to the proposed solution, you should raise it.   If this is just a matter of taste, that's out of scope.
[18:02:11] <Ole Troan> A significant change in the DHCP architecture, and adding IPv4 cruft in DHCPV6, that's hardly a matter of taste is it?
[18:02:41] 934ea03a4ea9ce8c leaves the room
[18:02:49] <Ted Lemon> What is the technical point that you are raising, Ole?
[18:03:03] Alex Petrescu joins the room
[18:03:08] <Ted Lemon> What will break if the working group follows the apparent consensus?
[18:03:18] <Ole Troan> First I would like to understand what problem you are trying to solve. Because I cannot make sense of what you are saying.
[18:03:33] <Ted Lemon> We are trying to deliver IPv4 configuration information to a node that is in an IPv6 island.
[18:03:47] <Ole Troan> You _can_ do what the working group propose, the state machine of creating dependencies between IPv4 and IPv6 provisioning is going to be tricky, but it probably is anyway.
[18:04:02] <Ole Troan> What I don't understand is why you want to make significant new work for a dying protocol.
[18:04:09] <Ole Troan> But I can make those points during the IETF last call.
[18:04:26] <Ole Troan> As I don't think the two of us will make much progress in this discussion.
[18:05:01] Dan York leaves the room
[18:05:11] <Ted Lemon> What I want to know is why you are so persistent in continuing this argument, that you would threaten an appeal over it.   What is it that you think the working group is doing wrong?
[18:05:29] <Ted Lemon> I get that you don't like the solution, but that's not a good enough reason for an appeal.   Everybody winds up in the rough sometimes.
[18:05:34] <Ole Troan> Not considering applying the existing architecture to the problem.
[18:05:50] <Ted Lemon> We did consider it.
[18:05:54] <Ted Lemon> Just now.
[18:05:58] <Ted Lemon> And for the past three IETFs.
[18:06:17] Andrew Yourtchenko leaves the room
[18:06:25] <Ole Troan> Well, please enlighten me then, cause I don't understand the arguments. And above you seemed to not want to discuss the technical issues.
[18:06:38] <Ole Troan> With regards to how IPv4 provisioning is tied to IPv4 transport.
[18:06:43] <Ted Lemon> No, I'm happy to discuss the technical issues.   You haven't raised a techincal issue.
[18:07:05] <Ole Troan> Nice of you to be the judge of that.
[18:07:21] <Ted Lemon> Ole, in order for it to be a technical issue, you need to say X breaks because of Y.
[18:07:33] <Ted Lemon> You haven't said any such thing.
[18:07:36] <Ole Troan> Where did you get that definition?
[18:07:46] <Ole Troan> I think this is decaying into a meta discussion.
[18:08:01] <Ole Troan> Let's do this at a different venue… be it IETF last call or somewhere else.
[18:08:17] <Ted Lemon> I don't know why that would be a better context, but whatever.
[18:08:32] <Ole Troan> What I'm proposing doesn't require new protocol specifications, and if people want to implement it, then the specs are all there.
[18:08:33] Andrew Yourtchenko joins the room
[18:08:47] <Ted Lemon> Ole, what you are proposing requires a change to the RFC2131 state machine.
[18:08:53] <Ole Troan> Which one?
[18:08:58] <Ole Troan> Which change that is?
[18:09:02] <Ted Lemon> There's no RENEWING state anymore.
[18:09:05] <Ted Lemon> Only REBINDING.
[18:09:18] <Ole Troan> RFC5107 deals with forcing all traffic through relay.
[18:09:31] <Ole Troan> So there is no direct server to client that needed port aware DHCP.
[18:09:45] <Ole Troan> Which means all clients can continue to use 67/68 as usual
[18:09:58] <Ole Troan> There is a force renew or something isn't there? I don't think that would work though. But is it used?
[18:10:25] <Ted Lemon> Forcerenew is used pretty much exclusively in ISP communications.   I don't know if it would be wanted in this case.
[18:10:47] <Ted Lemon> 5107 is an interesting point; I'm not sure if existing clients implement it, though.
[18:10:57] <Ole Troan> Well, I suppose it could be used if the DHCP server kept the option 82 information in the binding, but I haven't thought about it.
[18:11:03] <Ole Troan> 5107 would be relay only
[18:11:12] <Ole Troan> and the relay would have to be changed.
[18:11:31] Andrew Yourtchenko leaves the room
[18:11:40] <mcharlesr> Ted, how can I find out more about this IESG pushback on CGA use.  Would that be a DISCUSSes on a document, or was it the result of an earlier review of some kind?
[18:11:53] Andrew Yourtchenko joins the room
[18:12:14] <Ted Lemon> It was actually a point that Russ raised after the discussion was over.
[18:12:58] <Ted Lemon> His point is that CGA addresses shouldn't necessarily be recorded in lease databases.
[18:14:12] <Ted Lemon> In retrospect I'm not convinced that I agree with him, but it's water under the bridge at this point, and one advantage of the current approach is that it doesn't require CGA.
[18:17:56] <Brian Haberman> The DISCUSS and COMMENT points raised during IESG review are available at http://datatracker.ietf.org/doc/draft-ietf-dhc-secure-dhcpv6/ballot/
[18:18:19] <Brian Haberman> That does not cover the after-the-fact comments made by Russ.
[18:18:35] <Ted Lemon> I didn't find the text of Russ' DISCUSS helpful, hence the subsequent discussion.
[18:18:46] <Ted Lemon> Actually it might have been a comment.
[18:22:10] ogud leaves the room
[18:24:14] ogud joins the room
[18:27:15] ray.bellis joins the room
[18:27:54] Mark Townsley joins the room
[18:28:00] ray.bellis leaves the room
[18:37:40] Alex Petrescu leaves the room
[18:45:03] Mark Townsley leaves the room
[18:46:10] ogud leaves the room
[18:46:23] ogud joins the room
[18:47:10] ogud leaves the room
[18:47:47] <Bernard Volz> sheng is now presenting stateless reconfiguration.
[18:52:53] ogud joins the room
[18:58:24] Myung Hee Choi leaves the room
[19:10:33] Brian Haberman leaves the room
[19:32:10] Ian Farrer leaves the room
[19:33:10] ogud leaves the room
[19:33:36] Jinmei Tatuya leaves the room
[19:34:20] Andrew Yourtchenko leaves the room
[19:34:35] Ted Lemon leaves the room
[19:34:52] Suzanne Woolf leaves the room
[19:35:09] Suzanne Woolf joins the room
[19:35:18] Suzanne Woolf leaves the room
[19:37:40] pebersman leaves the room
[19:39:28] tomtaylor leaves the room
[19:43:34] mcharlesr leaves the room
[19:44:12] Suzanne Woolf joins the room
[19:45:30] Suzanne Woolf leaves the room
[19:51:03] Bernard Volz leaves the room
[19:53:25] danny leaves the room
[20:03:42] mcharlesr joins the room
[20:03:56] Suzanne Woolf joins the room
[20:11:20] mcharlesr leaves the room
[20:17:23] ogud joins the room
[20:19:02] mcharlesr joins the room
[20:31:55] mcharlesr leaves the room
[20:33:00] mcharlesr joins the room
[20:33:17] sarikaya2012 leaves the room
[20:34:29] ogud leaves the room
[20:38:49] mcharlesr leaves the room
[20:38:52] mcharlesr joins the room
[20:41:38] danny joins the room
[20:44:24] danny leaves the room
[20:47:30] Suzanne Woolf leaves the room
[20:47:45] Suzanne Woolf joins the room
[20:47:56] Suzanne Woolf leaves the room
[21:03:31] Ole Troan leaves the room
[21:05:02] Suzanne Woolf joins the room
[21:06:32] Jinmei Tatuya joins the room
[21:07:47] Jinmei Tatuya leaves the room
[21:18:20] Suzanne Woolf leaves the room
[21:18:25] Ian Farrer joins the room
[21:26:09] mcharlesr leaves the room
[21:29:30] Ian Farrer leaves the room
[21:31:43] Ted Lemon joins the room
[21:32:14] Ted Lemon leaves the room
[21:35:25] mcharlesr joins the room
[21:36:40] Ralf Weber joins the room
[21:36:59] Ralf Weber leaves the room
[21:43:45] pebersman joins the room
[21:43:49] pebersman leaves the room
[21:46:32] Mark Townsley joins the room
[21:48:02] Mark Townsley leaves the room
[21:50:45] Mark Townsley joins the room
[22:18:26] mcharlesr leaves the room
[22:19:29] mcharlesr joins the room
[22:26:36] mcharlesr leaves the room
[23:19:29] Andrew Yourtchenko joins the room
[23:45:10] Mark Townsley leaves the room
[23:58:18] Ralf Weber leaves the room
Powered by ejabberd Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!