[10:11:49] damien.saucez joins the room [10:11:57] damien.saucez leaves the room [12:13:39] ljakab joins the room [18:12:15] florin coras joins the room [18:19:52] mccreary joins the room [18:40:42] mccreary leaves the room [19:41:00] mccreary joins the room [19:46:19] gigix73 joins the room [19:52:04] shikob joins the room [19:54:16] mccreary leaves the room [19:56:51] Shane Amante joins the room [19:56:53] hartmans@mit.edu/barnowl joins the room [19:57:16] per@nordu.net joins the room [19:58:29] mccreary joins the room [20:01:05] Shane Amante leaves the room [20:01:36] Shane Amante joins the room [20:01:48] rdv@jabber.postel.org joins the room [20:01:58] arifumi joins the room [20:02:30] wmhaddad joins the room [20:02:41] mohsen joins the room [20:02:44] trond joins the room [20:03:06] ljakab leaves the room [20:04:03] damien.saucez joins the room [20:05:17] Lorand joins the room [20:05:41] Wes joins the room [20:06:29] agenda slide [20:07:56] draft-ietf-lisp-06 [20:08:00] fred templin at the mic [20:08:46] sam at mic [20:10:01] Scott Brim joins the room [20:11:28] mccreary leaves the room [20:11:38] HannuF joins the room [20:11:55] mccreary joins the room [20:13:28] draft-ietf-lisp+alt-03 [20:13:34] slide 4 [20:14:14] aalain joins the room [20:16:53] slide 5 [20:18:08] sriram at the mic [20:23:01] slide 3 [20:25:03] lixia joins the room [20:25:26] chris liljenstolpe joins the room [20:26:14] niall joins the room [20:27:03] rababy joins the room [20:28:33] lixia leaves the room [20:29:38] rababy leaves the room [20:30:21] dave meyer at the mic [20:30:37] sam [20:31:00] jari [20:32:48] moving on to "how to send Map-replies preso from next session [20:33:19] slide 2 problem statement [20:34:00] mccreary leaves the room [20:34:42] slide 3 af combinations [20:34:49] mccreary joins the room [20:34:53] rdv@jabber.postel.org leaves the room [20:37:57] slide 3 is a building slide, i believe we're on the 3rd iteration [20:38:40] should be full slide now [20:38:42] Can anybody send the slides out or make them available somewhere? [20:42:50] tetsuya.innami joins the room [20:42:52] chris - not sure if you're listening to the audio... [20:42:55] I am. [20:43:02] sorry.... [20:43:13] not your problem. [20:43:14] damien.saucez leaves the room [20:43:25] I'll try to scribe the slides that aren't diagrams [20:43:29] thank's. [20:43:42] well, once sam gets out of my way, that is ;-) [20:44:04] :) [20:45:02] tetsuya.innami leaves the room [20:47:42] mohsen leaves the room [20:47:53] slide 4 proposed solution [20:48:18] itr provides a set of itr-rloc addresses [20:48:33] ok [20:48:34] etr can select one of them for the map-reply destination addresses [20:48:45] etr no longer uses map-request source for map-reply destination [20:50:18] Simon Leinen joins the room [20:51:05] current map-request format.... [20:51:17] it's a format screen, you'd have to go to the draft [20:51:42] proposed map-request format, same drill [20:52:18] tetsuya.innami joins the room [20:53:29] slide 7 procedural [20:53:35] objections to putting this to -07? [20:53:40] needs to come to the list [20:54:50] per@nordu.net leaves the room [20:54:51] HannuF leaves the room: Computer went to sleep [20:55:04] moving to Memebership test for mapping info optimization [20:55:15] the propblem we are addressing [20:55:26] draft-flinck-lisp-membertest-00 [20:55:33] slide 1 [20:56:04] if an itr acting as a gateway between two namespaces doesnt' ahve a mapping for the destination eid, it needs to resolve the eid-to-rloc m,apping form a mapping system possibly usinga map resolveer [20:56:14] eids are allocated in blocks [20:56:39] some sites maybe too small to be worth a prefix allocation and related admin [20:57:04] damn, he's flying through some very text-heavy slides [20:57:22] the first packet without a eid-trolc mapping is dropped, buffered, or relayed over the mapping system [20:57:33] membership test [20:57:53] the auth etr provides to idr a membership test that represnts all or a subeset of the eid it can route to [20:58:17] membership test is an efficient compressed data struct to represent members that beloiing to the same set =- ETR [20:58:29] etr can express with it - the most frequesntly req. mappings [20:58:36] any other subset or all of the mappings [20:58:46] membership test doesn't require that the eids aggregate into prefixes [20:58:57] message handling with membership test [20:59:01] packet arrives from ahost to an itr [20:59:10] eid to rloc mapping cache look up [20:59:28] if the lookup doesn't result in a RLOC than a membershp test cache is checked [20:59:44] if any of the membership tests results into a match then the corresponding rloc is used [20:59:59] map-req message is piggybacked with the user packet to the rloc [21:00:09] if no matching membership test, use the map resolution [21:00:19] right column: [21:00:26] map-req message arrives to the auth ETR [21:00:34] slide changed... [21:00:55] message format - extends the lisp map-reply message format by defining a new mapping protocol data option [21:01:08] [packet diagram] [21:01:09] mccreary leaves the room [21:01:16] calculation of the membership test [21:01:27] niall leaves the room [21:01:28] mccreary joins the room [21:01:32] standard blook filter, defined by error probability [21:01:36] Thank's alot for doing this wes. [21:01:36] number of hash [21:01:43] test vector size [21:02:07] we fix the membership test for 2 types of small sites/enterprises, those with 100+ hosts, those w 1000+ hosts [21:02:30] the length of the vector tells the ITR which parameters to use [21:02:43] error prob = 10e-6 [21:02:48] hashes = 20 [21:02:55] m = 4096 or 32768 bits [21:03:00] hash functions are md5 [21:03:19] eid-prefixes shhould not be added to membership tests as this adds to the processing complexity [21:03:39] error handling and secuirty [21:03:53] niall joins the room [21:03:55] false-positive error case degenerates into stale mapping info case of baseline LIP [21:03:56] lisp [21:04:18] itr can trust an eid-to-rloc mapping from an etr, ti can trust the membership test asx well, no new security issues [21:04:25] niall leaves the room [21:05:42] that's it... [21:05:45] Wes leaves the room [21:05:53] Thank's Wes [21:06:01] shikob leaves the room [21:06:12] gigix73 leaves the room [21:06:13] arifumi leaves the room [21:06:21] tetsuya.innami leaves the room [21:06:31] Dapeng Liu joins the room [21:07:07] Shane Amante leaves the room [21:07:49] Lorand leaves the room [21:08:19] mccreary leaves the room [21:08:19] niall joins the room [21:08:29] florin coras leaves the room [21:08:34] niall leaves the room [21:09:22] wmhaddad leaves the room [21:12:09] Shane Amante joins the room [21:13:21] jeff.byzek@gmail.com joins the room [21:13:33] Shane Amante leaves the room [21:14:43] trond leaves the room [21:14:51] arifumi joins the room [21:19:37] arifumi leaves the room [21:28:48] chris liljenstolpe leaves the room [21:29:45] jpc joins the room [21:29:49] jpc leaves the room [21:31:16] chris liljenstolpe joins the room [21:52:55] Dapeng Liu leaves the room [21:57:06] Simon Leinen leaves the room [22:01:43] mcharlesr joins the room [22:04:57] Scott Brim leaves the room [22:09:24] chris liljenstolpe leaves the room [22:29:19] Scott Brim joins the room [22:29:54] Scott Brim leaves the room [22:46:10] hartmans@mit.edu/barnowl leaves the room [22:46:31] mcharlesr leaves the room [23:23:05] jeff.byzek@gmail.com leaves the room