[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <YsXcAKWZYR90JMFf@itl-email>
Date: Wed, 6 Jul 2022 15:01:19 -0400
From: Demi Marie Obenour <demi@...isiblethingslab.com>
To: oss-security@...ts.openwall.com
Subject: Re: Re: DO NOT OPEN PREVIOUS MAIL Re:
Denial of service in GnuPG
On Wed, Jul 06, 2022 at 04:50:11PM +0200, Solar Designer wrote:
> On Wed, Jul 06, 2022 at 09:47:28AM -0400, Demi Marie Obenour wrote:
> > On Wed, Jul 06, 2022 at 03:38:10PM +0200, Solar Designer wrote:
> > > On Wed, Jul 06, 2022 at 07:02:59AM -0400, Demi Marie Obenour wrote:
> > > > Was adding compression to PGP even a good idea in the first place?
> > >
> > > I think actually yes, it was, especially back then. It has probably
> > > helped more than it hurt in PGP's lifetime so far.
> >
> > Interesting. Why do you say that?
>
> Oh, I didn't feel this even needed explanation, and I feel silly writing
> the below and don't really have time for it (lesson re-learned: should
> have stayed silent), but well.
>
> PGP is commonly used on compressible data (such as text), and PGP
> messages are then transferred over a network and finally stay in
> people's mailboxes or such. Bandwidth was commonly low back then, and
> storage much more limited than today's.
>
> Some compression existed for unencrypted messages - some network links
> somewhat compressed (e.g., V.42bis), some mail clients supported mailbox
> compression, and of course a mailbox could also be compressed manually.
>
> Obviously, already encrypted content is not compressible.
>
> Without built-in compression in PGP, its messages would be slower to
> transfer and larger to store. Compression would need to be performed
> before PGP, which would be an inconvenience and would lead to similar
> risks, especially if automated, and would often not be done. In PGP,
> it's just one standard way to do it, not more than one.
>
> So compression was of some benefit to a lot of people. We could argue
> that it's little benefit, but multiplied by the number of people it's
> significant. Was compression also a problem for a lot of people?
> Theoretically, yes, but in practice those attacks were not common.
>
> We could also argue that PGP never became popular, MUA integrations are
> poor, etc., and as a consequence that its individual features were not
> of a lot of benefit to computer users at large. While true, that
> argument also means the risks associated with those features did not
> apply to most computer users. So it's irrelevant.
>
> What I say is that for the geeks using PGP, compression was overall of
> more benefit than risk.
>
> Oh, and I'm also grateful for compression in SSH, despite of my own
> criticism of its effect on security.
>
> Alexander
Thanks! I had not considered this at all, and it makes a ton of sense.
Being able to compress data prior to encryption is indeed necessary if
one wants to avoid using more space than a message compressed using
normal methods.
--
Sincerely,
Demi Marie Obenour (she/her/hers)
Invisible Things Lab
Download attachment "signature.asc" of type "application/pgp-signature" (834 bytes)
Powered by blists - more mailing lists
Please check out the
Open Source Software Security Wiki, which is counterpart to this
mailing list.
Confused about mailing lists and their use?
Read about mailing lists on Wikipedia
and check out these
guidelines on proper formatting of your messages.
