Raph Levien writes: # Earlier, I mentioned that two and a half protocols survived the # day. The remaining one is MSP. It's actually not a bad protocol. It appears to have been designed by the NSA, so that's not surprising in some senses. The question is, I think, how much baggage does it bring that's not really relevant for civilian/commerical use ? Debate about the use of sensitivity labels has recently resurfaced on the IPSEC list, although opinion seems to be running quite heavily in favor of implementing them (at the network layer) so far. Bill Stewart writes:
Where can we find the new specs for MSP?
With some help from Howard Weiss of Sparta Secure Systems Eng. (in MD), I finally found an online version. It's in 5 parts, accessible from http://bbs.itsi.disa.mil:5580/T3563 (look for MIL-STD-2045-18500). The web site is set up so as to make it a major pain to cut-and-paste or easily remember the precise URLs. They're zipped WordPerfect files, so I can't read them. If someone constructs a copy in ASCII or PostScript or HTML or something, let me know. According to http://www.itsi.disa.mil/dodiis/sec2-62.html, you can read about MSP in one of the Secure Data Network System (SDNS) Key Management Documents (NISTIR 90-4262). MSP is apparently (supposed to be) used in the DMS (Defense Message System). I think you can order hardcopy of such things from NIST. I found an archive of old traffic from the pem-dev list about a "Preliminary" MSP at: http://www.eff.org/pub/Privacy/Security/Crypto_misc/dod_pmsp_sdns.standards -Lewis "Shit !" -Pres. Richard M. Nixon, 1973