CURRENT_MEETING_REPORT_ Reported by Jeff Schiller/ MIT Mintues The SNMP Authentication Working Group met at the Pittsburgh IETF meeting on May 2, 1990. The primary focus of the meeting was a discussion of the relative merits of various Cryptographic Checksum algorithms used to ensure origination authentication and integrity of Protocol Data Units (PDUs). This discussion was the result of comments received from members of the Privacy and Security Research Group which reviewed the documents. Basically the problem boiled down to identifying which algorithms were both secure enough and yet were fast enough for the potential high traffic volumes that they may be needed to process. The algorithms discussed were: QMDC4, QMDC1, MD2, MD4, SNEFRU2, SNEFRU4. It was announced at the meeting that SNEFRU2 had been broken, and the consensus was that it therefore should not be considered. There was a sense that we needed to get cloture on the issue of what algorithm to use, in time for implementations to be demonstrated at Interop in October. Therefore the following decisions and action items resulted: o Consensus was reached that the RFC should *not* provide a menu of choices for implementors. Instead the RFC should specify just one of the candidate algorithms as the selected algorithm. This was argued on the basis that if more then one was allowed, each vendor would pragmatically need to support all of them, at a cost in terms of the development time for product, and memory size of the runtime binary. o Jeff Mogul and Chuck Davin volunteered to get performance numbers on the various candidate algorithms and post their results to the mailing list. The hope here is that of all the algorithms, sufficient number would be of high performance that at least one could be found that would be both fast and secure enough to pass a review by people who can judge the security of these types of algorithms. o The above work would be completed and a selection made in time to advance the three documents for consideration as "Proposed Standards" of the Internet. Since the meeting was held, the performance measures have been made and 1 it appears that MD4 is the clear performance winner. The documents will be changed to reflect this and submitted to the IETF with the recommendation they be progressed to the Proposed Draft state. ATTENDEES Hossein Alaee hossein_alaee@3com.com Stan Ames sra@mbunix.mitre.org Douglas Bagnall bagnall_d@apollo.hp.com Pat Barron pat@trqnsarc.com Pablo Brenner Alison Brown alison@maverick@osc.edu Ted Brunner tob@thumper.bellcore.com Jeff Carpenter jjc@unix.cis.pitt.edu Martina Chan mchan@mot.com Steve Crocker crocker@tis.com James Davin jrd@ptt.lcs.mit.edu Frank Kastenholtz kasten@interlan.interlan.com Louis Mamakos louie@trantor.umd.edu Keith McCloghrie sytek!kzm@hplabs.hp.com Jeffrey Mogul mogul@decwrl.dec.com Oscar Newkerk newkerk@decwet.dec.com John O'hara johara@mit.edu Brad Parker brad@cayman.com ? Mike Patton map@lcs.mit.edu David Perkins dave_perkins@3com.com Tod Pike tgp@sei.emu.edu Jonathan Saperia saperia%tcpjon@decwrl.dec.com Greg Satz satz@cisco.com Jeffrey Schiller jis@athena.mit.edu Richard Smith smiddy@dss.com ? Ted Soo-Hoo soo-hoo@dg-rtp.dg.com Michael StJohns stjohns@umd5.umd.edu Louis Steinberg louiss@ibm.com Ian Thomas ian@chipcom.com David Waiteman djw@bbn.com Steve Waldbusser sw0l@andrew.cmu.edu Y C Wang 21040 Homestead Rd Cupertino,Ca 95041 2