Another interesting article from Jeremy Allison (he had written earlier about the document lockup with MS Office 2007)
" ... Sure, there's gobs and gobs of extra software in the process which is usually run at the consumer end of the deal, trying to obfuscate and hide the fact that the consumer possesses all the information needed to decrypt the file they've just been given. They have to have been given this, else they can't listen to the song or watch the movie. Claiming that this process can ever be made secure from the people you've just given all this information to is like believing you can create a secure bank vault by drawing chalk lines on the pavement, piling the money inside and asking customers to "respect these boundaries". The media industries are trying to sell what they consider to be valuable data without any means of prohibiting access to it. This is not a business model that is ever going to work."
http://news.zdnet.com/2100-9588_22-6189011.html?tag=nl.e550
Another interesting article from Jeremy Allison (he had written earlier about the document lockup with MS Office 2007)
Good Article!
I especially like the analogy with the South Park Underpants Gnomes.
* Step 1: Create a DRM system. * Step 2: ??? * Step 3: Profit!
:D
Regards,
- vihan
On Fri, 2007-06-08 at 01:07 +0530, Praveen A wrote:
Another interesting article from Jeremy Allison (he had written earlier about the document lockup with MS Office 2007)
<snip>
would like to add one pearl of wisdom :P
Just like encryption is breakable. Yes, RSA with 4096bit key is very much breakable. Only problem is that we dont have enough computing resources to break such keys.
Similarly, it'll take time for "DRM" to mature enough be "full proof".
Now, dont flame me. I'm not in the pro-DRM camp but given enough time, engineers can, well, do anything silly :P
'nuff said :)
On 10-Jun-07, at 7:43 PM, Dinesh Joshi wrote:
much breakable. Only problem is that we dont have enough computing resources to break such keys.
who is we
On Sun, 2007-06-10 at 19:51 +0530, Kenneth Gonsalves wrote:
On 10-Jun-07, at 7:43 PM, Dinesh Joshi wrote:
much breakable. Only problem is that we dont have enough computing resources to break such keys.
who is we
we is you and me and the rest of the 6+ billion humans ;)
On 10-Jun-07, at 9:18 PM, Dinesh Joshi wrote:
On 10-Jun-07, at 7:43 PM, Dinesh Joshi wrote:
much breakable. Only problem is that we dont have enough computing resources to break such keys.
who is we
we is you and me and the rest of the 6+ billion humans ;)
in other words it is unbreakable
On Sun, 2007-06-10 at 21:26 +0530, Kenneth Gonsalves wrote:
who is we
we is you and me and the rest of the 6+ billion humans ;)
in other words it is unbreakable
Not forever. At a time DES was considered to be unbreakable. But now it is very much breakable and that too in a practical amount of time. Similarly, RSA will become breakable as the technology advances.
Unbreakable is actually the wrong word to use for encryption schemes. RSA is breakable even today but it'll take an exponentially large ( read: practically not viable ) amount of time to break. So it is as good as "unbreakable" - at the moment.
The only truly unbreakable encryption scheme is one time pad :)
On 6/10/07, Dinesh Joshi dinesh.a.joshi@gmail.com wrote:
Not forever. At a time DES was considered to be unbreakable. But now it is very much breakable and that too in a practical amount of time. Similarly, RSA will become breakable as the technology advances.
Not sure I agree with the flat analogy. The weakness of DES was its key length -- 56 bits was just right for NSA's supercomputers to crack the code in a reasonable amount of time back then. 64 bits would have made it much stronger.
Unbreakable is actually the wrong word to use for encryption schemes. RSA is breakable even today but it'll take an exponentially large ( read: practically not viable ) amount of time to break. So it is as good as "unbreakable" - at the moment.
One route to breaking a crypt is through algorithm flaws. An open algorithm will get fixed faster in such a case.
In normal cases, even if computing speed increases by 2 times every year, all I need to do is increase my key length by a bit to make my crypt much stronger than required to offset the increase in computing speed..
The most feasible way to break a crypt is to attack the weakest link -- users.
On 6/11/07, Siddhesh Poyarekar wrote:
On 6/10/07, Dinesh Joshi wrote:
Not forever. At a time DES was considered to be unbreakable. But now it is very much breakable and that too in a practical amount of time. Similarly, RSA will become breakable as the technology advances.
Not sure I agree with the flat analogy. The weakness of DES was its key length -- 56 bits was just right for NSA's supercomputers to crack the code in a reasonable amount of time back then. 64 bits would have made it much stronger.
Nope. The DES algorithm cannot be made more "secure". Any attempts at changing it like increasing the key length, increasing the number of permutations, cycles etc.. etc.. just weaken it. So the algorithm cannot be optimized further. Hence, Triple DES was invented :) It has been mathematically proved that by doubling the key length of the DES algo, doesnt actually "double" the security it provides. Rather it just simply remains the same.
One route to breaking a crypt is through algorithm flaws. An open algorithm will get fixed faster in such a case.
We're not debating about Open or Proprietary algorithms. Long back itself it was understood that security by obscurity is useless :)
In normal cases, even if computing speed increases by 2 times every year, all I need to do is increase my key length by a bit to make my crypt much stronger than required to offset the increase in computing speed..
Check out quantum computing :)
The most feasible way to break a crypt is to attack the weakest link -- users.
Stating the obvious? :)
It has been mathematically proved that by doubling the key length of the DES algo, doesnt actually "double" the security it provides. Rather it just simply remains the same.
It would certainly make a brute-force attack much harder. The "thinking machine" does brute-force attacks remember :)
On Tue, 2007-06-12 at 04:54 +0530, Parijat Garg wrote:
It has been mathematically proved that by doubling the key length of the DES algo, doesnt actually "double" the security it provides. Rather it just simply remains the same.
It would certainly make a brute-force attack much harder. The "thinking machine" does brute-force attacks remember :)
Get your facts straight before you reply. a message encrypted with a 128 bit key DES is no more secure than a 64 bit key DES. Actually the length of the key is only 56 bit in the latter case.
Double, tripling or quadrupling the key length wont increase the amount of work that a cracker will have to do!
On 6/13/07, Dinesh Joshi dinesh.a.joshi@gmail.com wrote:
Get your facts straight before you reply. a message encrypted with a 128 bit key DES is no more secure than a 64 bit key DES. Actually the length of the key is only 56 bit in the latter case.
Please cite a source for this. I'd like to know how DES is done with a 128 bit key.
The Data Encryption Standard (DES) implies the use of a 56 bit key. Change the key length and you'll have to change multiple parameters of the algorithm. For example, with a 56 bit key, the algorithm works on 64 bits of data at a time. This chunk definition, substituter blocks and permutation blocks will all have to change, hence changing the very nature of the algorithm. You will hence have to end up with a DES-like algorithm rather than the DES algorithm, albeit with a larger key size. Difficulty in brute forcing a crypt is then just a function of the key size -- more the bits, more difficult it is.
Here's apparently the first case of DES being broken by brute force:
http://www.interhack.net/pubs/des-key-crack/
PS: I'm not a crypto expert, I'm just trying to reason out stuff based on facts I read up. Please correct wherever applicable.
On 6/13/07, Siddhesh Poyarekar siddhesh.poyarekar@gmail.com wrote:
On 6/13/07, Dinesh Joshi dinesh.a.joshi@gmail.com wrote:
Get your facts straight before you reply. a message encrypted with a 128 bit key DES is no more secure than a 64 bit key DES. Actually the length of the key is only 56 bit in the latter case.
Please cite a source for this. I'd like to know how DES is done with a 128 bit key.
The Data Encryption Standard (DES) implies the use of a 56 bit key. Change the key length and you'll have to change multiple parameters of the algorithm.
@Siddhesh: You're right. The two basic steps - substitution and permutation - are designed carefully for 64 bit blocks. Increasing it for higher number of bits will be non-trivial in that guaranteeing its security is difficult.
@Dinesh: By definition, a brute-force attack is just cycling through all the possible keys. Therefore, in general, greater the bits in the key, the more combinations that require to be tested and hence harder the brute-force attack. This is independent of DES or 3-DES or any other key-based encryption algorithm.
Parijat
On Thu, 2007-06-14 at 00:37 +0530, Parijat Garg wrote:
@Dinesh: By definition, a brute-force attack is just cycling through all the possible keys. Therefore, in general, greater the bits in the key, the more combinations that require to be tested and hence harder the brute-force attack. This is independent of DES or 3-DES or any other key-based encryption algorithm.
Pfleeger. It has the whole ago and in layman's terms.
Double DES is as secure as DES with 57bit key while Triple DES does give more security.
On 6/14/07, Dinesh Joshi dinesh.a.joshi@gmail.com wrote:
Pfleeger. It has the whole ago and in layman's terms.
Double DES is as secure as DES with 57bit key while Triple DES does give more security.
True, and 3DES provides an effective key length of about 108 bits. But you still haven't cited your 128 bit DES example. It's not a jab; I couldn't find it on googling, which is why I asked.
On Thu, 2007-06-14 at 21:03 +0530, Siddhesh Poyarekar wrote:
True, and 3DES provides an effective key length of about 108 bits. But you still haven't cited your 128 bit DES example. It's not a jab; I couldn't find it on googling, which is why I asked.
Not 108 bit but 112bit. I already cited the source - Pfleeger. On the 128bit key issue, it'll definitely increase the work but that still doesn't make the algorithm more complex as DES works best with a 64bit key. Any attempt at changing it has always resulted in reducing the effectiveness of the encryption.
Dinesh Joshi wrote:
Just like encryption is breakable. Yes, RSA with 4096bit key is very much breakable. Only problem is that we dont have enough computing resources to break such keys.
How do you know that the NSA doesn't have the computing power already and that they're not reading all GPG encrypted mails passing via the USA?
They've done it before you know :)
-- Anant
-
Anant Narayanan -
Dinesh Joshi -
Kenneth Gonsalves -
Parijat Garg -
Praveen A -
Siddhesh Poyarekar -
Vihan Pandey