Followup -- encrypted hard drive enclosure for $139

Peter Gutmann pgut001 at cs.auckland.ac.nz
Fri May 23 21:24:15 PDT 2003


Declan McCullagh <declan at well.com> writes:

>The TDES engine offers the same throughput as that of DES engine as we have
>engineered the pipeline design such that performance isn't degraded on TDES.
>The only difference is the start up latency. The start up latency for X-Wall
>SE-64A is 90 nano second and for X-Wall SE-128A is 270 nano second.

The suspicion when this was discussed after the RSA conference was that, based
on performance figures, they were doing something like using 3 DES engines and
pushing 3 different blocks through them at once (or a single DES engine with
per-round pipelining), this restricting them to the insecure ECB mode.  Using
one of the parallelisable modes would have been a better move.  The fact that
they were using ECB lead to the inevitable suspicions that there'd be other
security problems present as well - I think Lucky's condemnation may have been
a bit strong (it's not totally insecure), but the use of ECB mode doesn't
exactly inspire confidence in the design of the rest of the system, even if it
was done for pragmatic reasons.

Peter.





More information about the cypherpunks-legacy mailing list