Timeline for Design of Initialisation Vector
Current License: CC BY-SA 3.0
11 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Nov 3, 2013 at 5:28 | history | edited | user10211 | CC BY-SA 3.0 |
deleted 25 characters in body
|
| Dec 3, 2012 at 0:37 | comment | added | Gilles 'SO- stop being evil' |
You seem to have the misconception that entropy is a rare resource that each generated random bit consumes. This is wrong. If your system has too little entropy, your random bits are no good, even if you don't make many of them. If your system has enough entropy, you're good for billions of billions of bits. So make sure to provision your systems with some entropy (if they're VMs, the host can inject entropy; how depends on your VM software and deployment method), and read from /dev/urandom, not from /dev/random.
|
|
| Dec 1, 2012 at 19:39 | vote | accept | Simon G. | ||
| Dec 1, 2012 at 15:13 | answer | added | Thomas Pornin | timeline score: 5 | |
| Dec 1, 2012 at 9:37 | comment | added | Simon G. | Thanks for the link. The answer appears to be don't worry about design and just use lots of random bits? My application is designed to run on server clusters. Every server is an exact duplicate of the others. There is no user input on any of them. The boot-up process is totally deterministic. Where is any entropy going to come from? It is a well known problem in such cases that /dev/random is empty and stays empty, forcing the use of the /dev/urandom. Because of this reality, I'd like to avoid a complete reliance on random numbers. | |
| Dec 1, 2012 at 9:30 | comment | added | CodesInChaos | The standard solution is to simply draw 128 random bits from a crypto PRNG and be done with it. | |
| Dec 1, 2012 at 2:06 | comment | added | David Cash | @makerofthings7: You might be interested in this paper (and references) eprint.iacr.org/2012/251.pdf | |
| Dec 1, 2012 at 1:16 | comment | added | makerofthings7 |
I wonder if there any mathematical proof or statistic that covers the uniqueness and randomness of /dev/random or equivalent for hardware-based random number generators?
|
|
| Dec 1, 2012 at 1:02 | history | tweeted | twitter.com/#!/StackSecurity/status/274679760855134208 | ||
| Dec 1, 2012 at 0:34 | review | First posts | |||
| Dec 1, 2012 at 0:40 | |||||
| Dec 1, 2012 at 0:15 | history | asked | Simon G. | CC BY-SA 3.0 |