Re: "incrementing" ecb mode

From: Ernst Lippe (ernstl-at-planet-dot-nl_at_ignore.this)
Date: 10/19/03 

Date: Sun, 19 Oct 2003 22:11:11 +0200

On Sat, 18 Oct 2003 00:11:39 +0200, Eris Pluvia wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Peter Fairbrother:
>
>> Are there any security problems with using a sort-of incrementing ecb
>> mode, like this:
>>
>> 1) Every file is encrypted with a different starting key.
>>
>> 2) Each block in a file is encrypted with an incremented key, ie each
>> block's key is 1 more than the previous block's key.
>
> My option is rehashing the key each block.

I assume that you mean that the key for a block is the hash of the
key of the previous block. This make it very expensive to update
just a few blocks in the middle of the file.

> The loss of performance is less
> than this caused by rescheduling.

Do you mean to say that the extra overhead due to hashing
is low compared to the overhead that is caused by by rescheduling?
(If I understand you correctly, you still propose to reschedule
the key for each block.)

greetings,

Ernst Lippe

Quantcast