Re: Encryption Pipe stage

[Jim Bohnsack]

I may have offered this in 2002, but I don't remember.  The FCOPY
PACKAGE from the IBM download page has a "GARBLE" option. 
FCOPY, itself, is an extremely useful tool especially with the ability to
create "FCOPY PACKLIBs".  The contents, not only can have
duplicate member names, but the compression really squeezes the gas out
of the input.
Here's the help for GARBLE:
GARBLE       Will compress data within
the file,  and "garble" the data using a 
            
user supplied encryption key in such a manner that the file cannot 
            
be reconstructed without  specifying the proper encryption 
key. A 
            
read is issued  for the 8 character  key or the 
password  can  be 
            
be specified via the PASSWORD
option.                             

Jim

At 10:49 AM 6/28/2006, you wrote:
In the almost 4 years since this
last reply, has anyone worked on an
encryption/decryption stage for PIPELINES? Or a library of callable
subro=
utines?
/Tom Kern
/U.S. Dept. of Energy
/301-903-2211

On Tue, 17 Sep 2002 10:55:37 +0100, Simon Knights
<[EMAIL PROTECTED]> wrote:
>Date
sent:             
Thu, 12 Sep 2002 18:18:27 EDT
>Send reply to:         
VM/ESA Discussions <VMESA-L@LISTSERV.UARK.EDU>
>From:                  
Aria Bamdad <[EMAIL PROTECTED]>
>Subject:               
Encryption Pipe stage
>To:                    
VMESA-L@LISTSERV.UARK.EDU
>
>> Hi,
>>
>> Does anyone know if there is a Pipe stage or Rexx code that will
do
>> encryption?
>>
>> Thanks in advance.
>
>Hi Aria,
>
>I know this reply is coming late and that you have already had a
number =
of
other replies.
>
>A couple of years ago I wrote a pipeline stage that would do
encryption.=
 I
decided to use
>the Blowfish algorithm as designed by Bruce Schneier of Counterpane
(see=
>
http://www.counterpane.com/blowfish.html). The reasons for
choosing
Blowfish were:
>
>1 - It is unpatented and Royalty free.
>2 - Can use a key from 32 to 448 bits.
>3 - Is faster than DES or IDEA.
>4 - (so far) has not been cracked.
>5 - Is reasonably straight-forward to implement.
>
>Basically its a symmetric block cypher that uses a 16 round Feistel
Netw=
ork
that uses 64
>bit blocks (ie; it uses private keys - there is no public key). As
usual=
,
Key management
>presents its own problems and this technique is really only useful
where=
the key does
>not change frequently.
>
>Rexx was ideal for prototyping the stage - was quick to develop and
test=
.
But to be of
>any practical use I had to rewrite it in Assembler - and this runs
about=
400 times faster
>that the Rexx version. It uses 2 input streams - one supplies the
key, t=
he
other is the
>data to be encrypted - which can be a stream of data of any length,
so y=
ou
can encrypt
>anything from a word to a large file of data. There is one output
stream=
.
>
>But downside to this reply, is that the code is not mine (though I
wrote=
it) but belongs to
>my employer and I am not permitted to distribute it. But maybe I have
be=
en
able to
>suggest to you an approach that works.
>
>Best wishes,
>
>Simon Knights.
>UK.
>=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D

Jim Bohnsack
Cornell Univ.
(607) 255-1760