Date: Thu, 28 Feb 91 16:56:35 -0500 From: John T Kohl <jtkohl@MIT.EDU> To: krbdev@MIT.EDU 0. next meeting is next Thursday, 7 Mar 1991, at 3pm in E40-316. 1. status reports: jtkohl: worked on replay cache stuff for safe/priv msgs used mprof to clean up various memory leaks played with ISODE to fix some leaks & other problems. tytso: finished up API conversion to standardize on the return style. working on kdb_util dump/load (salt changes) thinking about realm "quality" stuff jfc: still studying GSS interface, & considering user2user code jis: has been playing with SPX will be at meeting about GSS jon: () 2. GSS API stuff: will be a meeting next Tuesday with DEC folks about solidifying the GSS API code. we're all invited. 3. build space is tight again; we can probably squeeze more out of afsreq. 4. All files should eventually be converted to have the phrase "All Rights Reserved" right after the Copyright line, and if you edit the file, be sure that 1991 appears as one of the years in the copyright notice. Also please remove all #include's of <krb5/copyright.h>. 5. alpha/beta testing issues. we're not getting much in the way of useful feedback from testers, so we won't add any more to those we have now. We are about ready for a general distribution beta-test. We're waiting for: completion of principal salting stuff completion of subsession key stuff ISODE 6.8 test/update 6. discussion of subsession keys, key mixing, etc. Opinion boils down that: mixing keys is sufficiently non-portable among encryption types that we probably shouldn't provide an interface at the kerberos level. For applications that care to use separate subsession keys for one/both direction, the client side can have the mk_req_* routine select and return one such key; the server can ask the mk_rep routine to select one as well. 7. other random distribution questions: we should do some sort of code/protocol auditing, to increase our confidence in the system. discussion of where to draw the lines for things considered security-critical (read-side routines, KDC, encryption layers, maybe others?) discussion of beta-test rewards for those finding security bugs in the library. Action items: Jeff will talk to DEC about the export issues with SPX and what they've done w.r.t. distributions Jeff will keep eyes/ears out for any "standardization" of string2key algorithms. John K. will find a way to get DER out of ISODE 6.8. John C. will produce a user2user draft interface spec and circulate it for review.