NSIS Meeting Minutes -------------------- * Interop Report **************** Presented by Roland Bless Lars: I was told that the implementations are somewhat related. Roland: Implementations are different. The Roke Manor implementation is not publically available. Jukka has another new implementation. We may need to work on more test cases. Lars: For proposed standard there is already a high number of implementations available. Martin: Are the test programs available for free? Christian: Yes, they are available in source code. The implementations are different and hence I do not understand Lars's statement. Please send me comments regarding the test cases. Jukka: University of Helsinki is working on a C-based implementation that is again different. Luis: We have a Java-based implementation and it is different than the other implementations even from an architecture point of view. We have also found a lot of bugs in GIST in the past based on the design choices we made. * QoS NSLP ********** Presented by Jukka Manner Jukka reports about the WGLC comments. Everyone seems to be happy with the document. * NATFW NSLP ************ Presented by Hannes Tschofenig Hannes reports about the shortened security considerations section. * Mobility ********** Slides show a couple of open issues. No comments * RMD ***** Presented by Georgios Georgios goes through the most recent updates. Only a small set of changes. No comments * QSPEC ******* No presentations since there weren't any changes. Hannes points to the relationship to the DIME work on the Diameter QoS attributes and the dependency on the registry for * Y.1541 QOSM ************* Presented by Al Morton No presentation slides available. Only editorial changes. Al: Jerry retired. Roland: Is that a problem? Al: Jerry retired but he will continue his work on this document. Al will be the editor. Lars: I am OK with this. * NSIS Operation over IP tunnels ******************************** No presentation since the document has not changed. * State Machines **************** No presentation since the document has not changed. * GIST over SCTP **************** Presented by Christian Dickmann Christian: It it OK to specify the two MA-Protocol-IDs in a single document? Magnus: It is OK for me to specify them in one document. Christian: We could combine things loosely and hence we could use DTLS also with DCCP. Jukka: We have posted a comment a draft about DCCP. Hence it would be good to have it in separate documents. Robert: He does not know why you need a separate ID for DTLS. He will think about it. Magnus: There are differences between DTLS and TLS. I am also not sure about the security Christian: We haven't sorted that stuff out yet. Hence, we still have to think about it. Hannes was just volunteered to review the DTLS part of the document. * NSIS Multi-Homing Issues ************************** Presented by Hong Cheng Lars: I would like to push back on new items before the current items are finished. Hong: I would like to see it started but not at this point in time. Lars: There are no impacts to the design work. Hannes: I see them more as extensions to provide performance optimizations. Lars: Performance optimizations for a non-deployed protocol are useless. Hannes: I agree. * HyPath ******** Presented by Luis No comments * Inter-Domain Aggregation for the QoS NSLP ******************************************* Presented by Roland Bless Roland explains the problems he sees for inter-domain reservations. Hannes: Question whether the high-level architecture is shown somewhere. Is this related to the RMD QOSM work given that you extend the RMD region from intra-domain to inter-domain. This is interesting stuff. Georgios: I agree with Hannes. This is interesting work. * Authorization Tokens ********************** Presented by Jukka Manner Hannes: Motivates the work and argues that it is important Luis: Started implementation work Roland: This is important work for QoS NSLP and the NATFW NSLP. This is a must once we get the current charter item done. Jukka: We did an implementation * Peering ********* Jukka Hannes has not read the draft and asks some clarifying questions. * Discussions ************* Lars: We need to get GIST used for something. Informational documents would be helpful that explain how RSVP users why to switch to NSIS.