IETF
alto
alto@jabber.ietf.org
Tuesday, November 5, 2013< ^ >
martin.thomson has set the subject to: ALTO WG http://tools.ietf.org/wg/alto/agenda
Room Configuration
Room Occupants

GMT+0
[00:38:45] Abhay Roy joins the room
[00:40:35] Abhay Roy leaves the room
[01:37:29] Michael Scharf joins the room
[01:38:04] Alessandro Amirante joins the room
[01:38:24] Paolo Saviano joins the room
[01:43:28] Wendy Roome joins the room
[01:43:38] Richard Alimi joins the room
[01:43:42] vijay.gurbani joins the room
[01:44:01] <Alessandro Amirante> Slide 1: ALTO WG th 88 IETF Vancouver, Canada Enr
[01:44:03] <vijay.gurbani> Meeting starting now.
[01:44:04] <Alessandro Amirante> Slide 2: Note Well
[01:44:09] <Alessandro Amirante> Current presenter: Chairs
[01:44:10] <Alessandro Amirante> Slide 2: Note Well
[01:44:30] <Alessandro Amirante> Slide 3: Agenda
[01:45:32] <Alessandro Amirante> Slide 4: WG Progress
[01:47:46] <vijay.gurbani> Jan Seedorf at mic
[01:50:34] <vijay.gurbani> That was Wolfgang Beck at the mic.
[01:50:48] <vijay.gurbani> Richard Yang at mic.
[01:53:19] <Richard Alimi> In the case where an ALTO Server returns the same CIDR prefix in multiple PIDs, I think the client should pick an arbitrary one
[01:53:28] <Richard Alimi> [feel free to say at mic]
[01:53:37] <vijay.gurbani> One sec, I will do so.
[01:54:54] <Richard Alimi> I would say it is an error, but the client should have a graceful fallback
[01:54:58] <Richard Alimi> (yep - i got it :) )
[01:55:39] <Alessandro Amirante> Presentation stopped
[01:57:04] <Alessandro Amirante> Slide 2: Status of draft-ietf-alto-deployments Ve
[01:57:09] <Alessandro Amirante> Slide 3: New Table of Content in -08 Many Changes
[01:57:22] <Alessandro Amirante> Current presenter: Vijay
[01:57:23] <Alessandro Amirante> Slide 3: New Table of Content in -08 Many Changes
[01:57:31] <Alessandro Amirante> Slide 4: Major Changes in -08 List Part 1
[01:57:54] <Michael Scharf> Note that there is (as usual) a significant lag on audio (at least for me)
[01:58:15] <Alessandro Amirante> @Michael: are you listening to the Meetecho audio?
[01:58:25] <Alessandro Amirante> is should be near teal time
[01:58:36] <Michael Scharf> No, I'll try
[01:58:53] <Alessandro Amirante> Meetecho is at www.meetecho.com/ietf88/alto
[01:59:06] <Alessandro Amirante> Slide 5: Major Changes in -08 List Part 2
[01:59:24] <Alessandro Amirante> Meetecho RTSP stream at: rtsp://meetecho.meeting.ietf.org/th8fx6o.sdp
[02:00:05] Michael Scharf (Alcatel-Lucent Bell-Labs) joins the room
[02:00:26] <Alessandro Amirante> Slide 6: Open Issue: “Monitoring ALTO” Origin
[02:01:22] Michael Scharf (Alcatel-Lucent Bell-Labs) leaves the room
[02:01:34] Michael Scharf (2) joins the room
[02:02:53] <Michael Scharf> Thx
[02:03:03] <Alessandro Amirante> Slide 7: Summary and Next steps
[02:05:34] Victor Kuarsingh joins the room
[02:05:45] <Michael Scharf> +1 on Vijay's statement
[02:06:40] <Wendy Roome> I agree -- it's pointless to standardize a client API
[02:08:01] <Michael Scharf> To Richard: Yes, the directorate wants a formal metric definition (e.g., like IPPM)
[02:08:16] <Alessandro Amirante> Presentation stopped
[02:09:11] <Alessandro Amirante> Slide 1: ALTO Topology Service: Use Cases, Requir
[02:09:25] <Alessandro Amirante> Current presenter: RIchard Yang
[02:09:26] <Alessandro Amirante> Slide 1: ALTO Topology Service: Use Cases, Requir
[02:09:54] <Alessandro Amirante> Slide 2: Overview     
  •  Goal:    
 provide
[02:11:03] <Alessandro Amirante> Slide 3: Use    
 Case    
 I:    
 Topology    
 Adv
[02:11:59] <Alessandro Amirante> Slide 4: 4
[02:12:22] <Alessandro Amirante> Slide 5: 5
[02:13:12] <Alessandro Amirante> Slide 6: 6
[02:14:42] <Alessandro Amirante> Slide 7: 7
[02:14:52] <Alessandro Amirante> Slide 8:      
 
[02:14:56] <Alessandro Amirante> Slide 9:      
 
[02:15:12] <Alessandro Amirante> Slide 10: 10
[02:15:42] <Alessandro Amirante> Slide 11: 11
[02:16:03] <Alessandro Amirante> Slide 12: Use    
 Case    
 I:    
 Example    
 Work
[02:16:09] <Michael Scharf> Clarification question for slide 12: ALTO does not really support multiple servers per ALTO client, right?
[02:16:30] <Michael Scharf> (There is some initial text on this in the deployment draft -08)
[02:16:31] <Alessandro Amirante> Slide 13: 13
[02:17:09] <Alessandro Amirante> Slide 12: Use    
 Case    
 I:    
 Example    
 Work
[02:17:25] <Alessandro Amirante> Slide 14: Use    
 Case    
 II:    
 Large    
 Flow
[02:17:32] <Alessandro Amirante> Slide 15: Large-­‐Flow    
 Scheduling     
 
[02:17:46] <Richard Alimi> I'm not sure what's intended by that question - a client can aggregate data from multiple servers, right?
[02:18:04] <vijay.gurbani> which question, Richard?
[02:18:13] <Richard Alimi> Michael's question
[02:18:25] <vijay.gurbani> You want me to mic it?
[02:18:26] <Michael Scharf> Yes - but as far as I know, we have not fully thought how to do this (but I might be wrong).
[02:18:40] <Michael Scharf> I tried to write something on this in deployment draft -08
[02:18:46] <Michael Scharf> (This is not for room)
[02:18:57] <vijay.gurbani> ok
[02:19:54] <Wendy Roome> Can't a client can contact N servers? I didn't think that doesn't require an extension.
[02:20:01] <Richard Alimi> In my implementation, I handled the case of multiple servers. Requires some thought about how figure out which server's information to use.
[02:20:06] <Alessandro Amirante> Slide 16: Large-­‐Flow    
 Scheduling     
 
[02:20:29] <Michael Scharf> Multiple servers raises discovery issues, as one example
[02:20:52] <Alessandro Amirante> Slide 17: Use    
 Case    
 III:    
 Two-­‐Level
[02:20:53] <Michael Scharf> This is mentioned in the deployment draft already, for what it is worth
[02:21:33] <Richard Alimi> Right. Agreed.
[02:21:34] <Richard Alimi> Thanks
[02:21:47] <Richard Alimi> Multiple servers is basically the use case for third-party discovery :)
[02:22:12] <Alessandro Amirante> Slide 18: Use    
 Case    
 III:    
 Two-­‐Level
[02:22:28] <Michael Scharf> Yes, but if the client is not multi-homed, what is the starting point for the discovery?
[02:22:38] <Michael Scharf> (all this not for the room)
[02:22:43] <vijay.gurbani> ok
[02:23:06] <Richard Alimi> Think of the ALTO Client as a bittorrent tracker
[02:23:07] <Alessandro Amirante> Slide 19: Use    
 Case    
 III    
 Example:    
 R3
[02:23:38] <Richard Alimi> bittorrent tracker uses third-party discovery to find alto server for the bittorrent clients
[02:24:03] <Richard Alimi> bittorrent tracker uses third-party discovery to find alto server for the bittorrent clients
[02:24:18] <Alessandro Amirante> Slide 20: Use    
 Case    
 III    
 Example:    
 R3
[02:24:20] <Michael Scharf> Yes, agreed, the bittorrent case is well understood and documented already. I am more concerned about Richard's use cases. But I have not fully thought of it.
[02:24:29] <Richard Alimi> Got it - okay.
[02:24:36] <Richard Alimi> I need to head out now...
[02:24:41] <Richard Alimi> See you all later.
[02:24:43] <vijay.gurbani> have fun
[02:24:51] <Richard Alimi> Have a drink for me.
[02:25:02] <vijay.gurbani> you bet.  why stop at 1/
[02:25:20] <Alessandro Amirante> Slide 21: Use    
 Case    
 III:    
 Example    
 Wo
[02:26:09] <Alessandro Amirante> Slide 22: Use    
 Case    
 IV:    
 SDN     
 
[02:26:43] <Alessandro Amirante> Slide 23: Use    
 Case    
 IV:    
 SDN     
 
[02:28:11] <Alessandro Amirante> Slide 24: Use    
 Case    
 IV:    
 SDN     
 
[02:28:14] <Alessandro Amirante> Slide 25: Use    
 Case    
 IV:    
 SDN    
 (Maple)
[02:29:16] <Alessandro Amirante> Slide 26: Use    
 Case    
 IV:    
 Example    
 Wor
[02:30:18] <Alessandro Amirante> Slide 27: Use    
 Case    
 V:    
 e2e    
 Joint
[02:31:27] <Alessandro Amirante> Slide 28: Use    
 Case    
 V:    
 Example     
 
[02:32:04] <Alessandro Amirante> Slide 29: Chosen paths considered both net and com
[02:32:33] <Alessandro Amirante> Slide 30: Use    
 Case    
 V:    
 Example    
 Work
[02:33:04] <Alessandro Amirante> Slide 39: Next    
 Step     
 
[02:35:21] <Alessandro Amirante> Presentation stopped
[02:35:25] Michael Scharf leaves the room: Replaced by new connection
[02:35:26] mscharf joins the room
[02:36:01] <Alessandro Amirante> Slide 1: ALTO for Querying LMAP Results draft-see
[02:36:10] Michael Scharf (2) leaves the room
[02:36:11] <Alessandro Amirante> Current presenter: Jan Seedorf
[02:36:11] <Alessandro Amirante> Slide 1: ALTO for Querying LMAP Results draft-see
[02:36:32] <Alessandro Amirante> Slide 2: Overview
[02:36:50] Michael Scharf joins the room
[02:38:25] <Alessandro Amirante> Slide 3: Overall LMAP Measurement Framework
[02:39:24] <Alessandro Amirante> Slide 4: draft-seedorf-lmap-alto-02
[02:39:39] <Alessandro Amirante> Slide 5: Example Use Cases
[02:40:34] <Alessandro Amirante> Slide 6: Useful ALTO Extensions for LMAP Use Case
[02:41:36] <Alessandro Amirante> Slide 7: Case study: Analyzing a large-scale data
[02:42:23] <Alessandro Amirante> Slide 8: Case study: Analyzing a large-scale data
[02:42:40] <Alessandro Amirante> Slide 9: Case study: Analyzing a large-scale data
[02:44:29] <Alessandro Amirante> Slide 10: Feedback / Discussion
[02:45:11] <Alessandro Amirante> Presentation stopped
[02:45:26] <Alessandro Amirante> Slide 1: ALTO VPN Service draft-scharf-alto-vpn-s
[02:45:29] <Alessandro Amirante> Slide 2: draft-scharf-alto-vpn-service-01 Require
[02:45:46] <Alessandro Amirante> Current presenter: Vijay
[02:45:47] <Alessandro Amirante> Slide 2: draft-scharf-alto-vpn-service-01 Require
[02:50:19] <Alessandro Amirante> Presentation stopped
[02:50:35] <Alessandro Amirante> Slide 1: What's the Impact of Virtualization to A
[02:50:43] <Alessandro Amirante> Current presenter: Zhen Cao
[02:50:44] <Alessandro Amirante> Slide 1: What's the Impact of Virtualization to A
[02:50:45] <Alessandro Amirante> Slide 2: Introduction
[02:51:33] <Alessandro Amirante> Slide 3: NFV and ALTO
[02:52:32] <Alessandro Amirante> Slide 4: Impact of Virtualized Endpoints to ALTO
[02:55:14] <Alessandro Amirante> Slide 5: Next Step
[02:55:18] <Alessandro Amirante> Slide 4: Impact of Virtualized Endpoints to ALTO
[02:55:23] <Alessandro Amirante> Slide 5: Next Step
[02:55:54] <Alessandro Amirante> Slide 3: NFV and ALTO
[02:58:41] <Alessandro Amirante> Presentation stopped
[02:58:53] <Alessandro Amirante> Slide 1: dra:-­‐seedorf-­‐cdni-­‐request
[02:59:02] <Alessandro Amirante> Current presenter: Jan Seedorf
[02:59:03] <Alessandro Amirante> Slide 1: dra:-­‐seedorf-­‐cdni-­‐request
[02:59:13] <Alessandro Amirante> Slide 2: ALTO    
 within    
 CDNI    
 Request    

[03:00:00] <Alessandro Amirante> Slide 3: How    
 can    
 mandatory    
 types    
 
[03:00:22] <Alessandro Amirante> Slide 4: High-­‐Level    
 Example    
 of    
 S
[03:01:11] <Alessandro Amirante> Slide 3: How    
 can    
 mandatory    
 types    
 
[03:02:14] <Alessandro Amirante> Slide 5: Advantages    
 of    
 using    
 ALTO    

[03:03:43] <Alessandro Amirante> Presentation stopped
[03:04:01] <Alessandro Amirante> Slide 1: ALTO Traffic Engineering Cost Metrics
[03:04:11] <Alessandro Amirante> Current presenter: Qin Wu
[03:04:11] <Alessandro Amirante> Slide 1: ALTO Traffic Engineering Cost Metrics
[03:04:12] <Alessandro Amirante> Slide 2: Recap.
[03:06:12] <Alessandro Amirante> Slide 3: Motivation
[03:07:25] <Alessandro Amirante> Slide 4: ALTO Cost metrics Extensions
[03:09:43] <Alessandro Amirante> Slide 5: JSON format example for metric “delay�
[03:09:57] Victor Kuarsingh leaves the room
[03:10:00] <Alessandro Amirante> Slide 6: Open issue- schedule of cost metric
[03:11:30] <Alessandro Amirante> Slide 7: Next step
[03:12:19] <Alessandro Amirante> Presentation stopped
[03:12:37] <Alessandro Amirante> Slide 1: ALTO Extension for Overlay Routing
[03:12:56] <Alessandro Amirante> Current presenter: Haibin Song
[03:12:57] <Alessandro Amirante> Slide 1: ALTO Extension for Overlay Routing
[03:12:58] <Alessandro Amirante> Slide 2: History
[03:13:05] <Alessandro Amirante> Slide 3: An real use case
[03:14:17] <Alessandro Amirante> Slide 4: New Service: Overlay Routing Cost
[03:15:12] <Alessandro Amirante> Slide 5: Example: request
[03:15:37] <Alessandro Amirante> Slide 6: Example: response
[03:16:50] <Alessandro Amirante> Presentation stopped
[03:17:07] <Alessandro Amirante> Slide 1: ALTO Extensions to Support Application a
[03:17:14] <Alessandro Amirante> Current presenter: Young Lee
[03:17:14] <Alessandro Amirante> Slide 1: ALTO Extensions to Support Application a
[03:17:29] <Alessandro Amirante> Slide 2: Scope
[03:19:09] <Alessandro Amirante> Slide 3: ALTO NET-APP context
[03:19:47] <Alessandro Amirante> Slide 4: Large Bandwidth Use Case <draft-bernstei
[03:20:29] <Alessandro Amirante> Slide 5: ALTO Query Mechanism that filters the re
[03:21:56] <Alessandro Amirante> Slide 6: ALTO Response Mechanism with Reduction o
[03:22:06] <Alessandro Amirante> Slide 7: Info Model
[03:22:12] <Alessandro Amirante> Slide 8: Encoding Example
[03:22:48] <Alessandro Amirante> Slide 9: Alto Response:
[03:23:21] <mscharf> Clarification question: Are for the use cases considered in this draft IP addresses as endpoint identifiers always sufficient?
[03:23:25] <Alessandro Amirante> Slide 10: Summary
[03:23:52] <vijay.gurbani> will ask...
[03:24:54] <Alessandro Amirante> Presentation stopped
[03:24:57] <mscharf> thx
[03:25:06] <Alessandro Amirante> Slide 1: PID Property Extension for ALTO Protocol
[03:25:15] <Alessandro Amirante> Current presenter: Richard Yang
[03:25:16] <Alessandro Amirante> Slide 1: PID Property Extension for ALTO Protocol
[03:25:17] <Alessandro Amirante> Slide 2: Mo-va-on     
  •  A    
 network    
 
[03:25:45] <Alessandro Amirante> Slide 3: PID    
 Proper-es    
 and    
 Network
[03:26:42] <Alessandro Amirante> Slide 4: PID    
 Proper-es:    
 Retrieval     
 
[03:27:29] <Alessandro Amirante> Slide 5: PID    
 Proper-es:    
 Adver-sement     

[03:28:02] <Alessandro Amirante> Slide 6: PID    
 and    
 Endpoint    
 Proper-es:
[03:28:53] <Alessandro Amirante> Slide 7: Seman-cs:    
 PID    
 Property    
 Value
[03:29:58] <Alessandro Amirante> Slide 8: Consistency    
 of    
 PID    
 Proper-es
[03:30:41] <Alessandro Amirante> Slide 9: Consistency    
 of    
 Endpoint    
 and
[03:31:05] <Alessandro Amirante> Slide 10: Discussion:    
 Proper-es    
 under    
 
[03:31:29] <Alessandro Amirante> Slide 11: Possible    
 Formal    
 Defini-on     
 
[03:32:21] <Alessandro Amirante> Presentation stopped
[03:40:09] vijay.gurbani leaves the room
[03:40:15] Paolo Saviano leaves the room
[03:40:45] Michael Scharf leaves the room
[03:40:48] Wendy Roome leaves the room
[03:40:51] Richard Alimi leaves the room
[03:41:10] Alessandro Amirante leaves the room
[03:41:35] Wendy Roome joins the room
[03:43:11] Wendy Roome leaves the room
[03:54:56] mscharf leaves the room
[06:42:00] Victor Kuarsingh joins the room
[07:21:35] Victor Kuarsingh leaves the room
[13:43:36] Victor Kuarsingh joins the room
[14:29:36] Victor Kuarsingh leaves the room