Re: Tasking and Concurrent Programming Using Multiple Ada Programs.

[Austin Obyrne <austin.obyrne@hotmail.com>]

On Saturday, April 29, 2017 at 2:56:37 PM UTC+1, Simon Wright wrote:
> Austin Obyrne <austin.obyrne@hotmail.com> writes:
> 
> > I have come over from sci crypt group to ask your advice on an
> > encryption concept.  The concept is a design plan to use a possible 36
> > ciphers (variants of the same basic cipoher and compatible to each
> > other)in an Ada tasking program.
> >
> > The hoped for plan is to write a program that will cater for 36 (or
> > less if needs be ) ciphers all reading in and immediately encrypting,
> > in rotation from the same file of plaintext and writing out the
> > resulting ciphertext to the same ciphertext file (the encrypted
> > plaintext file).  They would of course be written in Ada.
> 
> I dont know why you want to do this, but - if you want to get a speed-up
> - you won't. The problem isn't so much the reading of each successive
> plaintext character and sending it to the encrypting task; the problem
> is that the part of the program that writes the output must successively
> wait for each of the encrypting tasks to provide its latest output
> before writing it to the output medium. So the whole process is bounded
> by the output.
> 
> On the other hand, the effect of using tasks would be that each task
> would see only every Nth character, so perhaps this is actually your
> intent. If I was doing this I would look into using an abstract data
> type (say, Encryption_State) to encapsulate the state needed to encrypt
> the next character; arrange a procedure
> 
>    procedure Encrypt (State : in out Encryption_State;
>                       Plain : Plain_Text_Character;
>                       Result : out Encrypted_Text_Character);
> 
> 
> and then arrange
> 
>    type State_Index is mod 36;
>    States : array (State_Index) of Encryption_State;
> 
> so that as you pick the next input character you cycle round States to
> pick the state to encrypt it with.

Many , many thanks Simon.

I think the ciphertext bit of your advice will be OK.  Frankly I am not good enough to try out the Encrypt procedure.  what I am geting from you is tacit confirmation that the plan is at least feasible.

Speed of execution is not the target but the fact of each plaintext being encrypted by a different cipher is just unheard tactics todate and could be a new concept in cryptography.

Also, there are 36 factoorial permutations of key order that Alice might use and
exhaustive testing (Brute force per se) works out to about 1 million years (off the top of my head calculation.

The topic is fairly well covered in my Ada study book but I needed confirmation from you experts.

 A thousand thanks from me.

I will start experimenting when the remainder of the 36 variants is complete - done.

Austin O'Byrne