In Search for the Proper Crypto System
From: sammy (notvalid_at_it.com)
Date: 07/26/04
- Next message: Matthijs Hebly: "Re: "All random sci.crypt posters eventually exhibit periodicity"?????"
- Previous message: Francois Grieu: "Re: ISO 9796-2 2002"
- Next in thread: Bruce Stephens: "Re: In Search for the Proper Crypto System"
- Reply: Bruce Stephens: "Re: In Search for the Proper Crypto System"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
Date: Mon, 26 Jul 2004 12:33:10 -0600
Hello. I'm in the concept stage of a project that I think will require
cryptography as it's means to make it viable. However, I'm not a
cryptologist. My background is computer science and some AI. I could spend
months studying cryptology, but first I'd like to request your opinions as
to whether or not any crytographic algorithm or protocol or whatever can be
used to solve this problem. Then, I can expedite my research in those areas
you suggest. I sincerely thank you for your time and comments.
First of all, I can think of one solution to this problem, and that would be
the use of an asymetrical key pair. Unfortunately, this is not a practical
solution because of what I understand to be the speed limitations of using
an asymetrical key cryptology. I say that because I would not being using a
public/private key to encrypt only the symetric key used to encrypt the data
itself, but would use the asymetric public key to encrypt the data and then
the private key to decrypt it. If this could be done in a reasonable
timeframe on a 10MB data file, then that would be the perfect solution. A
reasonable amount of time would be about 2 minute for a 10MB file. Correct
me if I've been mislead here on the speed limitation of asymetric key use.
***Note: "public" and "private" are really misnomers in this system since
the private key is eventually revealed. The point here is that one key
encrypts, and another key decrypts!
I'll explain why a public/private key system will work (although not
practically because of speed) and if you could please tell me if there
exists an alternative crypto system or something like it or close to it
without the speed limitation, I would appreciate it and will be able to
focus my research on that area in order to test this concept of mine to see
if anything in cryptology can be used to make this work.
Imagine the source of a large data file d originates from an authority A and
sent to client C. Then, two or more of these clients exchange this same
data file.
A-->d-------> C1 ---d--> C2 ---d--> C3---d...--> Cn
Before d is sent to anyone, it is encrypted with the private key. That is,
before A sends it to the first client, C1, and before any client sends it to
the next client. A is simply the original source of d. However, before d
is encrypted, each client must obtain a new public key from A to encrypt it
with. The public/private key pair can be different each time d is
retransmitted.
Then, when the new client receiving encrypted d, it must obtain the private
key from A in order to decrypt it. So, from the diagram above, A encrypts d
with a public key and sends d to C1. When C1 is ready to read d, it request
A for the private key. Then, before C1 retransmits d to C2, it requests
another different public key from A to encrypt d before retransmission.
After d is encrypted again, it is sent to C2. Again, before C2 can decrypt
d, it must obtain the private key from A, and so on.
The point here is that, the clients may or may not be trustworthy. If
symetric keys were used instead, it is possible for for each client to pass
the key along with the data, d, to each client so they can avoid contacting
the authority, A, for the key each time to decrypt d. When using an
asymetrical key, each client can be directed to encrypt with a specified key
and the receiving client MUST contact A for the private key if they want to
read the data.
So, can this be practically done with asymetric keys? If not, as I suspect,
is there another area to focus my research that can do something similar?
Thank you very much,
Sam
- Next message: Matthijs Hebly: "Re: "All random sci.crypt posters eventually exhibit periodicity"?????"
- Previous message: Francois Grieu: "Re: ISO 9796-2 2002"
- Next in thread: Bruce Stephens: "Re: In Search for the Proper Crypto System"
- Reply: Bruce Stephens: "Re: In Search for the Proper Crypto System"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
Relevant Pages
|