Encryption increase possibility of corruption of file?

rinmanb70 wrote:

> I was wondering if anyone has seen that encryption increases the
> possibility of corrupting files. In other words, would anyone say
> before one encrypts important files, one should weigh the possibilty
> that there is an increased likelihood of losing the data due to
> corruption?
>


Encryption can be seen as a way of randomizing the data and adding
spurious data. So a stream of characters xxyzzyxx might turn
into !a@b#c$d&e which changes the number of runs of 0s and 1's and
makes the decryption of letters dependent upon the letters that came
before it! This latter situation means that you might recover
xxypdwaxx - in other words certain types of encryption cause a single
bit error to affect multiple bytes! Most good encryption, however,
includes more redundancy than the original so that decryption protects
your data. Error detection is as important as error recovery.

Use winzip (or similar tool) and stop worrying.

Rick Merrill

Sebastian Gottschalk <> wrote:

> rinmanb70 wrote:
> > Borked Pseudo Mailed wrote:
> >
> >> Obviously. Encryption certainly does increase the chances that some
> >> minor file corruption could make all your data inaccessible.
>
> Layering a level of error correction codes among a file has the same
> net effect for the encrypted version of the same file.
>
> Now could someone please abstract it? Encryption only obfuscates
> existing redundancy. Error correction based on reliance of existing
> redundancy is none, it's just poking around.

*laugh*

You MUST be related to Dale Beaudoin.

TwistyCreek

Rick Merrill wrote:
> rinmanb70 wrote:
>
> > I was wondering if anyone has seen that encryption increases the
> > possibility of corrupting files. In other words, would anyone say
> > before one encrypts important files, one should weigh the possibilty
> > that there is an increased likelihood of losing the data due to
> > corruption?
> >
>
>
> Encryption can be seen as a way of randomizing the data and adding
> spurious data. So a stream of characters xxyzzyxx might turn
> into !a@b#c$d&e which changes the number of runs of 0s and 1's and
> makes the decryption of letters dependent upon the letters that came
> before it! This latter situation means that you might recover
> xxypdwaxx - in other words certain types of encryption cause a single
> bit error to affect multiple bytes! Most good encryption, however,
> includes more redundancy than the original so that decryption protects
> your data. Error detection is as important as error recovery.
>
> Use winzip (or similar tool) and stop worrying.


What good is to tell me that "most good encryption" has built in
protection against corruption if you can't give me examples of those
that do and those that don't. Sounds like you just like to hear
yourself talk even though you don't have anything constructive to say.
And telling me to use Winzip and stop worrying sounds like you like to
think you know more than the next guy (you may, or in some cases, you
may not) and you like to make yourself look better by trying to make
other people look bad.
So, from now on, if you don't want to help, then stay out of my
questions.

rinmanb70

Hadron Quark <> wrote:

> "rinmanb70" <> writes:
>
> > Sebastian Gottschalk wrote:
> >
> >> Corruption and losing data are two different things.
> >
> > I probably wasn't clear in my question. What I'm asking is if
> > encrypting a file makes it slightly more likely over time that the
> > file will become corrupted versus if that file had never been
> > encrypted.
> >
>
> Define corrupted? if you mean any byte changing then why would it make
> any difference whether its encrypted or not? Encrypted files dont give
> off a special scent which attracts disk defects you know.

OK, to crawl out on the theoretical branch for a moment, I'd suggest
that certain types of encryption implementations just might do exactly
that.

Consider an encrypted container, and the fact that it has to be mounted
and unmounted for access to the rest of the data. This means a certain
"header" section is being accessed quite a bit more frequently than the
data contained inside, or "loose" data someone else on the disk. It
could be argued that MTBF statistics say the encrypted volume would
more likely to fail do to disk corruption than the areas of the platter
where that 10 year old email from "Mom" that you haven't read in 9
years resides.

George Orwell

George Orwell wrote:

> OK, to crawl out on the theoretical branch for a moment, I'd suggest
> that certain types of encryption implementations just might do exactly
> that.
>
> Consider an encrypted container, and the fact that it has to be mounted
> and unmounted for access to the rest of the data. This means a certain
> "header" section is being accessed quite a bit more frequently than the
> data contained inside, or "loose" data someone else on the disk. It
> could be argued that MTBF statistics say the encrypted volume would
> more likely to fail do to disk corruption than the areas of the platter
> where that 10 year old email from "Mom" that you haven't read in 9
> years resides.


I hadn't thought of this, but it seems very possible. 5 stars from me
for this answer George.

rinmanb70