Steve Schear (schear@lvcm.com)
Wed, 24 Jun 1998 13:25:48 -0700
At 1:52 PM -0400 6/24/98, Matt Blaze wrote:
>My own experience with CFS is that, given modern (486 and later) processors,
>multiple-megabyte memories, reasonable (cleartext-side) filesystem caching
>algorithms, and typical (Unix, at least) user access patterms, ANY modern
>cipher, even 3DES, is probably fine. I use CFS and 3DES for about half of my
>home directory (which is NFS-mounted from a shared server), and I frankly
>don't notice any difference between that and, say, SAFER (which CFS
>also supports). The encypted directories are slower, but not so much
>slower that I mind or especially notice.
>
>On the other hand, performance for an encrypted disk device driver might be
>another story altogether.
Will Price (now at PGP) did an excellent job, first with CryptDisk and
later with PGPDisk. These are both device drivers and have good
performance. Although both are for the Mac, the overall design decisions
should be applicable to most platforms.
PGPDisk has not gotten much publicity. Is the Windows version out yet?
Does anyone know if the source to either is available. I'm sure it could
be a boon to anyone considering this area.
--Steve
---------------------------------------------------------------------
reply to schear - at - lvcm - dot - com ---
PGP mail preferred, see http://www.pgp.com and
http://web.mit.edu/network/pgp.html
RSA fingerprint: FE90 1A95 9DEA 8D61 812E CCA9 A44A FBA9
RSA key: http://keys.pgp.com:11371/pks/lookup?op=index&search=0x55C78B0D
---------------------------------------------------------------------
The following archive was created by hippie-mail 7.98617-22 on Fri Aug 21 1998 - 17:19:00 ADT