Last Modified: 2004-09-07
Done | Submit requirements document to IESG | |
Done | Submit framework document to IESG | |
Nov 04 | Submit forwarding element functional model document to IESG | |
Mar 05 | Submit formal definition of controlled objects in functional model | |
Mar 05 | Submit protocol selection/definition document to IESG | |
Mar 05 | Submit applicability statement to IESG |
RFC | Status | Title |
---|---|---|
RFC3549 | I | Linux Netlink as an IP Services Protocol |
RFC3654 | I | Requirements for Separation of IP Control and Forwarding |
RFC3746 | I | Forwarding and Control Element Separation (ForCES) Framework |
Minutes ForCES WG at the IETF61 in Washington DC, USA
----------------------------------------------------- Agenda, IETF 61 Monday, November 8, 2004, 1300-1500 ==================================== CHAIRS: David Putzolu <David.Putzolu@intel.com>, Patrick Droz <dro@zurich.ibm.com> ADs: Alex Zinin <zinin@psg.com>, Bill Fenner <fenner@research.att.com> AGENDA: 5 min - WG status & Agenda Bash - chairs 30 min - ForCES Protocol Draft Status & Changes Jamal Hadi Salim draft-forces-protocol-01.txt 10 min - ForCES LFB Path Discussion Robert Haas on behalf of Weiming Wang (no draft available) 20 min - Options for LFB-level multicast and related issues Robert Haas (no draft available) 15 min - IP TML Proposal Alex Audu draft-audu-forces-iptml-00.txt 20 min - TIPC Transport Mapping Layer Proposal Jon Maloy draft-maloy-tipc-01.txt 5 min - ForCES IP TML Proposal Furquan Ansari draft-khosravi-tcptml-00.txt 15 min - ForCES Element Discovery Proposal Furquan Ansari draft-ansari-forces-discovery-01.txt 10 min - Flexinet EU ForCES Protocol Implementation Experience Robert Haas (no draft available) Attendees: 50 Minutes: The chairs began the meeting by reviewing the agenda and then going over the status of the group. It was stated the the protocol was expected to go through 1 or 2 more drafts before going to WG Last Call. Jamal gave a presentation on the protocol covering both the current structure and the open issues. He asked for comments on the open issues: - possibly adding a CE protocol LFB but that is still an open question - open issue on how path data is recorded. - controversial issues: -what is a path: - a map to a targeted entity - or 32 identifier similar to SMI OID - issue of how to access multiple rows - open question of whether content based access possible - packing/transport of path referenced data. Authors are looking for input on all the issues Q.- are they all in tracker - A. believe so, will check At the moment there are 8 open issues on the on the list: http://www.mip4.org/issues/tracker/forces/ There was no further discussion on the protocol draft ............................................................... Robert presents on behalf on Weiming the status of the Path LFB path discussion. Joel - Weiming is very concerned that subscripting include explicitly what field being used as index. He has a valuable point but I am not sure I got it. By what is Weiming subscripting - regular subscript - content field Joel believes every table should have a regular subscript. Weiming has the opinion that we need to explicitly indicate what we are using for selection. Joel comment: assume we have specified what you are going subscript, but where is the subscript to be found? or is there another sequence of information that indicates where the subscript is found? Joel wonder whether one must always have explicit indices that are index based and content (field ) based. A path should always have attributes and the fields. Joel thinks semantically putting it in the data is the wrong place. We need a clear indication what the target of operation is. Jamal wants closure of the discussion. Robert is not sure he sees the difference. David - need to resolve over the next few weeks probably best is to setup a call with Weiming to discuss. .................................................................. Robert presented options for LFB-level multicast and related issues. The question is how does one set information into several LFBs. There are three ways of doing multicast: 1) merged multicast: use PL-level multicast for LFB multicasting 2) split multicast that needs both PL level and LFB level (proposal from Joel, Steven) 3) xcast that only has unicast PL messages, list of LFB explicitly included in message proposal from Weiming Q: can you multicast to all LFBs of a given class. A: as is no, but if it is important we should find a way to support it. .............................................................. Alex presents the TML draft. He starts with the TML functions and design criteria such as simplicity, statelessness, security and trust model. Joel is surprised by what he found in the TML draft. He is looking for how to really do it. Either select a protocol with a transport or use TCP with explanation of how to transport. A: Well it is an API Joel: we need to find the transport David: this needs to be added Joel: that is where they should start David; so you are talking about a service access point Jamal: I was going to write a draft on this Joel: a service definition has to exist jamal: do we expect to see two different vendors to provide a TML and a PL At this point the lengthy discussion of what this draft should be was taken off line. But further questions came: Joel: I am missing the semantics of the join message. Why do we need TML level join message. Critical piece are missing, what is carrying the message. Alex: what transport do you want to map to Joel: well what transport do you want to map to. You must pick what the transport is: Jamal: this is a first cut. Joel: but this should have come first. This discussion was again take off line. Q; why only one way authentication A: Feels that the CE should be trusted. Q: there are several different models Jamal: there are several different trust models Q: what if there is a CE spoofer, could take over the network A: but FE initiated so it should know who it is initiating to. ................................................................. John Malloy presents TIPC. It is a transport protocol connection less and connection oriented Reliable and unreliable including multicast. Can be reliable in one direction And unreliable in the other direction. The protocol does not have any security Features. It is assumed to be operated in a trusted environment. It does have congestion control. They have control and data connections which have lower priority. The main purpose of the draft is to promote TIPC as the TML for ForCES. ............................................................... Furquan presented a TCP/IP based TML for ForCES. Control and data connections are used. Security is provided through TLS. It also supports prioritization and it protects against DoS attacks. David: the draft needs more details explaining how it should be used. ............................................................... In a second presentation Furquan presented a draft on ForCES Element Discovery Proposal. Only very few people had read it so consideration as a WG was deferred to the list. The content of the draft falls under the current ForCES charter. ............................................................... Finally Robert presented the FlexiNET project. FlexiNET is a EU project that will make use of ForCES. The objective of FlexiNET is to accelerate the introduction of next generation marketable services, and increase the competitiveness in the telecom field by facilitating the broadening of current business model for services provisioning and exploitation. |