--- dmolnar <dmolnar@hcs.harvard.edu> wrote:
1) untrusted and "impersonal" 2) untrusted but personalized 3) trusted and "impersonal" 4) trusted and personal
"Trusted" in "devices" generally assumes tamper-proofness (so that, once certified, the trust is implied the next day, with some p close to 1.) Brain is on the top of that list, ms windows at the bottom. The top secrets reside in brain (passphrases), lower-grade secrets (passphrase-protected) on electrical gadgets. The problem is the capacity - both storage and computational - which drops as trustworthiness increases. I think that these tradeoffs are not well understood, too much is assumed. On average, effective passphrase length is less than 60 bits. Yes, it's hard to estimate what is the effective length of strings composed from personal experience and published bits, but I have seen no data that would support anything more than said 60 bits. So if 256-bit block cipher is at the end of the security chain tha begins with 60-bit passphrase, why, OH WHY do we bother ? I mean, even 56-bit single DES in ECB mode will stop all casual harvesting. We need to enhance storage capacity of the most trusted environments. This ultimately leads to implants, with keychain items as an in-between step. __________________________________________________ Do You Yahoo!? Get email at your own domain with Yahoo! Mail. http://personal.mail.yahoo.com/