| |
 |
|
|
Science Forum Index » Cryptography Forum » PGP...
Page 1 of 1
|
| Author |
Message |
| Amir Abdollahi... |
 Posted: Sat May 10, 2008 10:39 am |
|
|
|
Guest
|
--- In the PGP scheme, what is the expected number of session keys
generated before a previously created key is produced?
--- In PGP, what is the probability that a user with N public keys
will have at least one duplicate key ID? |
|
|
| Back to top |
|
| David Eather... |
| Posted: Sat May 10, 2008 8:32 pm |
|
|
|
Guest
|
Amir Abdollahi wrote:
Quote: --- In the PGP scheme, what is the expected number of session keys
generated before a previously created key is produced?
If the RNG is good - which it is normally assumed to be, and you are
using 256 bit keys (AES, Twofish) then 2^128 for CAST with a 128 bit key
2^64.
Quote: --- In PGP, what is the probability that a user with N public keys
will have at least one duplicate key ID?
Sub Key ID is 32 bits long, so the answer is 2^16 or 65536.
What you need to look up is "birthday paradox" which will tell you the
whys and wherefores. |
|
|
| Back to top |
|
| Tim Smith... |
| Posted: Sat May 10, 2008 9:11 pm |
|
|
|
Guest
|
In article <4YKdnbP4xoAE0bvVnZ2dnUVZ_o3inZ2d at (no spam) supernews.com>,
David Eather <eather at (no spam) tpg.com.au> wrote:
Quote: --- In PGP, what is the probability that a user with N public keys
will have at least one duplicate key ID?
Sub Key ID is 32 bits long, so the answer is 2^16 or 65536.
Uhm...should the probability be somewhere in [0,1]? :-)
--
--Tim Smith |
|
|
| Back to top |
|
| David Eather... |
| Posted: Sun May 11, 2008 5:38 pm |
|
|
|
Guest
|
Tim Smith wrote:
Quote: In article <4YKdnbP4xoAE0bvVnZ2dnUVZ_o3inZ2d at (no spam) supernews.com>,
David Eather <eather at (no spam) tpg.com.au> wrote:
--- In PGP, what is the probability that a user with N public keys
will have at least one duplicate key ID?
Sub Key ID is 32 bits long, so the answer is 2^16 or 65536.
Uhm...should the probability be somewhere in [0,1]? :-)
Doh! Yes of course.
If you have N keys, out of a total possible M keys, the chance of a match is
P = 1 - ((M-1)/M * (M-2)/M * ...(M-(N-1))/M)
which would simplify into something much prettier - maybe this (no
warranty expressed or implied - going from memory with no sleep for
30hrs). On the plus side, if I am wrong someone will correct me.
M!
P = 1 - ----------------
(M-(N-1))! x M^N |
|
|
| Back to top |
|
| |
|
Page 1 of 1
All times are GMT - 5 Hours
The time now is Fri Sep 05, 2008 12:35 am
|
|