IETF
grow
grow@jabber.ietf.org
Tuesday, November 11, 2014< ^ >
Rob Shakir has set the subject to: GROW - IETF90
Room Configuration
Room Occupants

GMT+0
[03:17:32] Markus de Bruen joins the room
[03:22:16] Markus de Bruen leaves the room
[03:23:08] Markus de Bruen joins the room
[03:30:38] Dan York joins the room
[03:31:06] Meetecho RAV joins the room
[03:32:39] Meetecho joins the room
[03:33:54] John Scudder joins the room
[03:35:07] <John Scudder> Introducing myself, your humble Jabber scribe. Are any of you actually remote?
[03:35:25] <John Scudder> Chris showing chairs' slides, on slide 2
[03:35:31] <John Scudder> draft status
[03:37:41] <John Scudder> Now Iljitsch
[03:37:46] Jeffrey Haas joins the room
[03:38:06] Jeffrey Haas has set the subject to: GROW - IETF91
[03:39:03] <John Scudder> slide 2
[03:39:04] <John Scudder> slide 3
[03:39:39] <John Scudder> slides flipping back and forth
[03:39:40] <John Scudder> slide 7
[03:39:51] <John Scudder> slide 8
[03:41:43] <John Scudder> slide 9
[03:42:18] <John Scudder> slide 10
[03:43:10] <John Scudder> slide 11
[03:44:02] <John Scudder> slide 12
[03:44:40] <John Scudder> slide 13
[03:45:26] <John Scudder> slide 14
[03:45:30] <John Scudder> slide 15
[03:45:37] <John Scudder> 16
[03:45:50] <John Scudder> 17
[03:46:08] <John Scudder> 18
[03:46:19] <John Scudder> 19
[03:46:23] <John Scudder> 20
[03:46:46] <John Scudder> 21
[03:47:58] <John Scudder> // greenwich meriden or paris meridien?
[03:48:02] <John Scudder> slide 22
[03:48:33] <John Scudder> 23
[03:48:38] <John Scudder> 24
[03:48:38] <John Scudder> 25
[03:48:40] <John Scudder> 26
[03:48:42] <John Scudder> 27
[03:49:30] <John Scudder> back to 24
[03:50:18] <John Scudder> 25
[03:50:32] <John Scudder> 26
[03:50:44] <John Scudder> slide 28
[03:51:26] <John Scudder> 30
[03:52:40] <John Scudder> ruediger volk
[03:53:20] Markus de Bruen leaves the room
[03:53:26] <John Scudder> questions that getting address blocks is expensive, although consultants are expensive
[03:55:29] <John Scudder> overall point (I think) don't try to band-aid bad practice, instead encourage users to move to good practice (PA space that is)
[03:56:08] <John Scudder> notes that bhutan telecom lacks footprint in germany
[03:56:56] <John Scudder> rejoinder is that it's handy to have one aggregate for easy firewall config
[03:57:03] <John Scudder> also, multihoming
[03:57:46] <John Scudder> cathy aronsen
[03:58:00] <John Scudder> geo addressing has never worked
[03:58:05] <John Scudder> and it won't work this time
[03:58:31] <John Scudder> Iljitsch
[03:58:39] <John Scudder> actually geo addressing has never been tried
[03:58:40] <John Scudder> cathy
[03:58:59] <John Scudder> it's not physically possible. the network isn't built like that. the topology doesn't actually follow geography.
[03:59:03] <John Scudder> joel j
[03:59:42] <John Scudder> again, topology is interesting, geography isn't
[04:00:02] <John Scudder> Iljitsch
[04:00:21] Cengiz  Alaettinoglu joins the room
[04:00:32] <John Scudder> (I think the point is that at a continental level there is some relevance to geography)
[04:00:36] <John Scudder> joel j
[04:00:54] <John Scudder> (I think the point is, no it's not. at this point I am going to stop transcribing yes/no/yes/no)
[04:00:57] <John Scudder> chris m
[04:01:05] <John Scudder> please stop ratholing
[04:01:16] <John Scudder> paraphrasing Iljitsch
[04:03:05] <John Scudder> "is providing guidance to an enterprise … a useful thing for GROW?"
[04:03:17] <John Scudder> specifics in draft we can deal with later
[04:03:19] <John Scudder> Iljitsch
[04:03:24] <John Scudder> example:
[04:04:17] <John Scudder> jen l
[04:04:57] <John Scudder> how is this different from v4 space?
[04:06:09] <John Scudder> Iljitsch
[04:06:23] <John Scudder> big v6 address blocks provide more scope for mischief
[04:06:34] <John Scudder> at least that is what I think he will get around to saying
[04:06:34] <Jeffrey Haas> (more aggregateable)
[04:07:27] <John Scudder> peter: last one
[04:07:30] <John Scudder> shin m
[04:07:58] <John Scudder> how am I supposed to designate my tokyo-san francisco link?
[04:08:01] <John Scudder> Iljitsch
[04:08:05] <John Scudder> you mean the p2p?
[04:08:10] <John Scudder> don't advertise that into bgp
[04:08:12] <John Scudder> read the draft
[04:08:29] <John Scudder> shin:
[04:08:39] <John Scudder> so my ip allocation design will be impacted by this?
[04:08:51] <John Scudder> peter: work on it some more and get back to us
[04:08:59] <John Scudder> Now Sriram presenting
[04:09:34] <John Scudder> slide 2
[04:12:09] <John Scudder> slide 3
[04:12:25] <John Scudder> "a proposed working definition of route leak"
[04:12:56] <Dan York> John Scudder - thanks for the great level of scribing.  I was in another room watching for when this presentation came up.  Your level of detail was very helpful.
[04:13:06] <John Scudder> glad to hear it!
[04:13:22] <John Scudder> slide 4
[04:14:42] <John Scudder> slide 5
[04:14:52] <John Scudder> (u-turn with more specific prefix)
[04:15:45] <John Scudder> slide 6
[04:16:35] <John Scudder> (hijack with data path to legit origin)
[04:16:39] <John Scudder> slide 7
[04:16:50] <John Scudder> (type 4: leak of internal prefixes and accidental deagg)
[04:18:13] <John Scudder> slide 8
[04:18:21] <John Scudder> (type 5: lateral isp to isp leak)
[04:19:46] <John Scudder> slide 9
[04:19:53] <John Scudder> (not claiming to be exhaustive)
[04:20:06] scott joins the room
[04:20:54] <John Scudder> final slide
[04:21:13] <John Scudder> asking chairs for WG adoption
[04:21:20] <John Scudder> q/c?
[04:21:39] <John Scudder> <mumbled name>
[04:21:51] <John Scudder> valley-free is already well-defined
[04:22:07] <John Scudder> the distinction between the types is whether the attacker modifies the path
[04:22:13] <Jeffrey Haas> (andrei robachevsky)
[04:22:15] scott leaves the room
[04:22:17] <John Scudder> thanks
[04:22:42] <John Scudder> sriram:
[04:23:13] <John Scudder> WG football — came up in SIDR, go to GROW, see if you can define a problem, then maybe IDR
[04:23:45] <John Scudder> andrei:
[04:24:54] <John Scudder> peter cuts mike lines
[04:25:03] <John Scudder> dan york
[04:25:13] <John Scudder> I found this and also the slides useful, thanks
[04:25:19] <John Scudder> brian dixon
[04:25:34] <John Scudder> to refute the comment about valley-free, there are examples that can be shown valley-free but still routeable
[04:25:52] <John Scudder> e.g. customer->customer's peer violates policy but dosn't have valleys
[04:25:56] <John Scudder> sriram:
[04:26:03] <John Scudder> brian has other cases he wants added
[04:26:06] <John Scudder> geoff huston:
[04:26:12] <John Scudder> valley-free doesn't encompass everything
[04:26:24] <John Scudder> it assumes ASes have consistent policy for all their prefixes. not true.
[04:26:36] <John Scudder> also assumes you can infer relationship from a distance
[04:26:53] <John Scudder> even lixia's original paper admits there is a lot of unclassifiable stuff
[04:27:21] scott joins the room
[04:27:21] <John Scudder> leaks that create different outcomes in forwarding are the ones that matter
[04:27:37] <John Scudder> #include relevant mr. spock quote
[04:28:27] <John Scudder> show of hands for adoption
[04:28:29] <John Scudder> many hands for
[04:28:35] <John Scudder> no hands agaist
[04:28:38] <John Scudder> against
[04:28:57] <John Scudder> chris: one more item not on agenda
[04:29:10] <John Scudder> robert and brian haberman work in sidr
[04:29:14] <John Scudder> robert from ripe ncc:
[04:29:49] <John Scudder> rpsl
[04:30:06] scott leaves the room
[04:30:08] <John Scudder> resurrecting draft in sidr, would like grow review, would like rpsl review
[04:30:12] Dan York leaves the room
[04:30:14] <John Scudder> meeting ends
[04:30:17] John Scudder leaves the room
[04:30:19] Cengiz  Alaettinoglu leaves the room
[04:31:42] Jeffrey Haas leaves the room
[04:33:00] Meetecho leaves the room
[04:33:00] Meetecho RAV leaves the room
[04:36:42] Jeffrey Haas joins the room
[04:37:19] Jeffrey Haas leaves the room
[04:46:22] Dan York joins the room
[04:46:27] Dan York leaves the room
[18:55:39] Jeffrey Haas joins the room
[19:17:41] Jeffrey Haas leaves the room