IETF 88 - Vancouver, BC Canada TUESDAY, November 5, 2013 (IETF 81) 0900 - 1130 : Plaza A -------------------------------- Spencer Shepler and Brian Pawlowski - co-chairs === Intro / Blue Sheets Spencer blue sheeted and noted well. === Internationalization - Dave Noveck Many musts is spec, but not done through RFC's bis, 3010bis04. stringprep'ed added after review. Implementations worked, but the spec was diverging. Dave would like us to read RFC's especially chapter 12. FedFS document - Spencer - being held in queue, let us know if dependency on 3530bis... IESG wants and may insist on a more limited/rational approach to stringprep. Implementations may restrict ideal - Martin says that IESG will likely accept existing implementations that work. See slides. Noveck suggests a new I18N separate manageable document. === RPCSEC GSS - Andy Adamson Simplify draft04 -> draft05. Removed anything not needed for 4.2. GSS3 handle, payload and server/client agreements. Nico wanted channel binding, compund authentication, and security labels. Andy has worked with Nico (original author). See slides. === Server side security - Andy Adamson Decipher the issues. Revolves around READ stateid being unique to client/user. Martin suggests a cryptographic hash vs. a "random number". Really really really see the slides. === Migration Update - Dave Noveck rfc3530. Updates for migration. See slides. === NFSv4 Migration experience - Chuck Lever Make sure prototyped before moving through the IETF process. Lease moves on migrations - Chuck is stepping through the ramifications. See slides for Chuck's recommendations on Transparent State Migration (TSM) === Multi-Domain Federated File System Requirements - Andy Adamson Andy covered constraints including security constraints. See slides. === Problems with minor versioning - Dave Noveck Problems both of feature bashing, implementation and fixing "bugs" taking longer than wanted. Original intent of minor versioning was small incremental features, but that promise has not panned out (latency to feature release has been huge). Spencer comments - individual feature sets may cater to single non-interoperable solutions. See slides. === Client Encryption - Marcus Watts See slides. === Flexible Files Layout draft - Benny Halevy A quick update - see slides. === Mark Eshel quick on XATTR Brief introduction of the discussion occurring within the NFSv4 implementation community for Linux and the possible desire to define xattr-like capabilities for NFSv4. Discussion time was running very short so the discussion would be carried to the WG alias.