Minutes for Rserpool - IETF 65 Introduction: Lyndon gave an introductory status report for the group. The group still has outstanding comments from GenArts review on the architecture, threats and comparison documents, and these still need to be resolved, esp. the architecture and security comments. We also need to proceed faster to get the protocol documents into last call. A hum from the participants established that most participants feel the protocol specifications are in good shape for last call, although later Randy noted that he had received some comments from implementers that could be added to further clarify some text. Reliability for IPFix: draft-bclaise-ipfix-reliability-01.txt Peter Lei Peter Lei gave a presentation on the use of Rserpool for IPfix. There is strong interest in the IPfix WG for using Rserpool to address reliability issues, especially for billing records. RFC 3917 - Requirements for IP Flow Information Export (IPFix) describe reliability requirements that are necessary for billing apps. This internet-draft shows how the Rserpool architecture fits well with the IPFix device(s) and collectors to satisfy the reliability requirements. Some issues have been raised by the IPFix group that we need to address. One is that they need additional Rserpool servicing policies. These include allowing them to specify a favored server, rather than round robin, for example. The policy would be, use the "main server" if it is available, otherwise any backup server is fine. A second policy that is of interest to them is if the "main server" comes back up after a service interruption, they want to go back to using the "main server" instead of the secondary ones. We need to look into adding these policies to Rserpool. Rserpool was designed to allow new policies to be added to the base set. This can be added to the policy internet draft draft-ietf-rserpool-policies-02.txt and will not impact advancing the main protocol docs. A second issue involves knowing the last billing record that was processed by the IPFix application before a server fails. We discussed handling this thru the Rserpool cookie mechanism. We will work further on these issues with the IPFix WG on these two issues and address them in the next draft. DTLS/SCTP: Michael Tuexen draft-tuexen-dtls-for-SCTP-00.txt Michael Tuexen presented some related security work for SCTP, involving the adoption of DTLS for SCTP to make implementation simpler and more efficient, and to allow security to work with SCTP extensions such as PR-SCTP. This draft will be implemented in OpenSSL. Moving the group ahead: We had a discussion with the AD and determined the following plan. The editors will review and revise all the documents (especially the architecture doc) that had pending IESG comments. Fred Baker would do a review for the WG of the complete set. We also need to last call the main protocol documents, ASAP, ENRP and Common parameters. If these reviews finds them ready then the set then goes back to the IESG. The schedule is for this to happen before IETF-66. Maureen emphasized the need to progress the protocol specs and reevaluate the architecture draft and finish off the work of the group.