Other Directory Sites: SeekWonder | Directory Owners Forum
The following archive was created by hippie-mail 7.98617-22 on Sat Apr 10 1999 - 01:18:25
EKR (ekr@rtfm.com)
01 Feb 1999 06:50:28 -0800
Mok-Kong Shen <mok-kong.shen@stud.uni-muenchen.de> writes:
> Jim Gillogly wrote:
> >
>
> >
> > M-K Shen wrote:
> > > Brad Kemp wrote:
> > >>
> > >> If there is any data loss (i.e. lost packet or corrupted disk) the
> > >> rest of the plaintext is lost.
> > >
> > > I have addressed this point in a response to Jim Gillogly. If
> > > communication line is functioning, their protocol should take care
> > > of lost packets and retrasnsmission. Certainly my scheme does not
> > > address hardware problems on the side of communication partners. But
> > > that is also not addressed by other schemes, if I don't err.
> >
> > Taking these in reverse order: You do. It is. True enough. Not
> > relevant to your session key scheme. Not adequately.
> >
> > 'Nuff said... on this end, anyway.
> >
>
> I don't understand. The above was addressed to Brad Kemp. It is
> certainly o.k. that you answered in his place. But your answer does
> not appear to have 'substance' and I don't yet see your answer
> to the other of my response that was addressed to you. BTW, my
> English is unfortunately not so good as to comprehend your last
> sentence well. Perhaps you would be kind engough to employ simpler
> plain English for correspondce to a foreigner.
To expand Jim's rather terse, though correct error:
1. You do err:
2. A large number of other session key generation schemes cope
with packet loss adequately and resynchronize. (IPSEC comes to mind.)
3. Your scheme doesn't address hardware problems on the side of
communications partners. (Incidentally, believing that this
is the only possible cause of data loss is monumentally naive.)
4. Retransmission isn't relevant (or at least shouldn't be) to
your session key generation scheme. Note that at least one
annoying consequence of your scheme is that even if you have
retransmission, your scheme exhibits pathological behavior
in the case of out-of-order delivery.
5. You haven't addressed Jim's original message adequately.
Tbe quite frank, your response to Brad indicates fairly clearly
that you aren't very familiar with much of the communications
security literature and as a result, you're proposing inferior
solutions to problems that are already fairly well solved.
Read the literature.
-Ekr
-- [Eric Rescorla ekr@rtfm.com]
The following archive was created by hippie-mail 7.98617-22 on Sat Apr 10 1999 - 01:18:25