Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > sci.crypt > #50735

Re: Implementation of the Rumba20 compression function

Show all headers | View raw

On 01.11.21 23:58, Ed Carp wrote:
> On Sun, 31 Oct 2021, space aliens made Leo write:
> 
>> I noticed Rumba20 [1] has been out for quite some time, but I've never
>> come across an actual implementation or deployment of it.
>>
>> Depending on the security / speed required, there are a lot of
>> parameters to tweak including
> 
> I guess I'm missing something - what does security have to do with 
> compression? I mean, if I want to compress data, I wouldn't really care 
> about anything else but compression ratios. Encryption would be a 
> separate step.
> 
> <scratching head>

It's a great exercise that I can appreciate. I bet Leo can tell you more 
about this, for now just my two cents:

Compression allows you to use a low entropy source (e.g. me thinking of 
a clever password) to generate a high entropy piece of data. You simply 
make the low entropy piece of data long enough so it contains the 
desired amount of entropy. Then, by compressing that data, that entropy 
can be condensed to a shorter piece of data (up to its maximum entropy 
of course and only if the compression function doesn't suck.).

Then again, there are proven methods for building a hash function from a 
collision resistant compression function (Merkle Damgard), then for 
building an HMAC from a hash function and finally a key derivation from 
an HMAC. One could keep this exercise going by making a CSPRNG form the 
HMAC (NIST Special Publication 800-90A) and I bet Chris would be willing 
to help implementing the Rumba20-HMAC into his symmetric HMAC-cipher.

So, in the end you got a small building block (the compression function) 
that you can derive from about every cryptographic tool you need (short 
of asymmetric encryption / digital signatures). But let's see what Leo 
has to say about this.

Cheers,

Max

Back to sci.crypt | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread

Thread

Implementation of the Rumba20 compression function Leo <usenet@gkbrk.com> - 2021-10-31 18:34 -0500
  Re: Implementation of the Rumba20 compression function Max <maxturv26@gmx.net> - 2021-11-01 01:13 +0100
    Re: Implementation of the Rumba20 compression function Leo <usenet@gkbrk.com> - 2021-11-01 03:42 -0500
  Re: Implementation of the Rumba20 compression function Ed Carp <ecarp@lsisoft.net> - 2021-11-01 17:58 -0500
    Re: Implementation of the Rumba20 compression function "Chris M. Thomasson" <chris.m.thomasson.1@gmail.com> - 2021-11-01 17:01 -0700
      Re: Implementation of the Rumba20 compression function Ed Carp <ecarp@lsisoft.net> - 2021-11-01 23:18 -0500
      Re: Implementation of the Rumba20 compression function Leo <usenet@gkbrk.com> - 2021-11-02 08:41 -0500
        Re: Implementation of the Rumba20 compression function Ed Carp <erc@pobox.com> - 2021-11-02 12:14 -0500
    Re: Implementation of the Rumba20 compression function Max <maxturv26@gmx.net> - 2021-11-02 01:37 +0100
      Re: Implementation of the Rumba20 compression function "Chris M. Thomasson" <chris.m.thomasson.1@gmail.com> - 2021-11-03 12:59 -0700
        Re: Implementation of the Rumba20 compression function Max <maxturv26@gmx.net> - 2021-11-03 21:13 +0100
          Re: Implementation of the Rumba20 compression function "Chris M. Thomasson" <chris.m.thomasson.1@gmail.com> - 2021-11-03 13:40 -0700

csiph-web