Re: Java Compression & Encryption

["Shakti Saran \(College of Computing\)" <snsaran@cc.gatech.edu>]

Thank you for your 0.02$. I am concerned about the compression algorithm
because compression algorithms tend to put header information (which is
standard) that can be used by the cryptanalyst to break the compressed and
encrypted data. I was hoping that if the Java community has come across any
compression algorithms (perhaps available open source) that combined with
encryption are less vulnerable to cryptanalysis. Or, if the community has
used any techniques to achieve the same.

Thank you again for your help.

Shakti Saran
Computer Science & Discrete Math
Georgia Tech

----- Original Message -----
From: <Vikrant.Verma@alltel.com>
To: <snsaran@cc.gatech.edu>; <ajug-members@ajug.org>
Sent: Friday, June 27, 2003 11:19 PM
Subject: RE: Java Compression & Encryption


> I wonder what this post is doing on  AJUG. What u really need is a crypto
newsgroup. Anyways let me share my 0.02$.
> I would suggest you to use PKZIP, the decision of this algo has nothing to
do with unicity but rather its the simplicity of implementation. Java
already has an implementation of PKZIP and it should not be hard for you to
figure out how to use it. Also unicity can be better answered if you use
combination of the following techniques compression, scrambiling, expanding
and substituting data.
> I noticed that you are a college student, so maybe basing my answer on
simplicity will not appeal to you. So let me explain how pkzip will work wrt
increasing unicity distance. Unicity is simply the minimum length of the
text that can be uniquely deciphered by trying brute force attack on a large
encrypted message. This can help an attacker to find the key used to encrypt
the text. This kind of an attack depends highly on readability of the
message  Once the data is compressed the readability of the message is
reduced. A compressed data will always have a higher unicity and the unicity
distance will always be proportional to compression. On the hindsight the
higher the compression the faster the brute force attack, if the attacker
can guess the compression algorithm. Compression factor of any non lossy
algorithm is directly dependent on the data set. So an algorithm which is
good for a data set may not be as effective for another. Therefore if i may
generalize, any compressi!
> on algorithm will serve the purpose and pkzip is easily available.
>
> -Vikrant
>
> PS: Software Rule No 1: KISS
>
> -----Original Message-----
> From: Shakti Saran (College of Computing) [mailto:snsaran@cc.gatech.edu]
> Sent: Fri 6/27/2003 4:08 PM
> To: ajug-members@ajug.org
> Cc:
> Subject: Java Compression & Encryption
>
>
>
> I am trying to build a small application that will read from the database,
compress and encrypt the information and send it out as an email attachment.
I am using Rijndael for security. I want to compress information inorder to
increase the unicity distance. I need help in selecting a good compression
alogrithm that will do the above.
>
> Thank you for the help.
>
> Shakti Saran
> Computer Science & Discrete Math
> Georgia Tech
>
>
>