Managing entropy in computing devices for cryptographic key generation
This disclosure describes cryptographic secure implementation of a Pseudo Random Number Generator (PRNG) architecture based on existing Fortuna algorithm, but providing improvements thereupon for gathering and measuring entropy. The improvement includes a unique step of initial seeding that is not covered by Fortuna. The solution should be adapted to a variety of computing and communicating devices, including mobile devices.
1. A computer-implemented method of managing entropy for secure cryptographic key generation in a computing device, the method comprising:
initializing a seeding process by adding an initial amount of data to an entropy pool; harnessing entropy by starting entropy daemons to gather entropy from a plurality of entropy generating sources in the computing device to add to the entropy pool; and managing amount of entropy in the entropy pool by ensuring that there is a minimum amount of entropy in the entropy pool while sending out respective outputs responding to corresponding incoming requests for data, thereby maintaining a required security strength of the cryptographic key; wherein the entropy pool is one of a plurality of entropy pools, and the N-th pool of the plurality of entropy pools is reseeded for every 2 N -th incoming request for data; wherein the method includes using encryption algorithms; wherein the encryption algorithms integrate cross-platform accumulation of entropy into the entropy management scheme.
2. The method of claim 1 , wherein the managing step is performed by one or a combination of the following: managing queue of incoming requests, managing reseeding of the entropy pool, and managing length of the output data stream.
3. The method of claim 2 wherein the managing step further includes: deciding which entropy pool to use in response to an incoming request.
4. The method of claim 2 , wherein reseeding is performed by hashing the specific entropy pool.
5. The method of claim 2 , where each entropy pool collects enough entropy between reseedings to ensure the required security strength of the cryptographic key.
6. The method of claim 1 , wherein the initial amount of data added to the entropy pool in the initializing step comprises at least 64 bytes of data.
7. The method of claim 1 , wherein the initialization step comprises: loading a seed file into the entropy pool.
8. The method of claim 7 , wherein the seed the has a seed state of a cryptographically secure pseudo random number generator (CSPRNG) written on an area of a memory of the computing device that is read on application start up.
9. The method of claim 8 , wherein the seed file is backed up at least once, wherein a current seed state is available on the next application startup, obviating the need to re-establish entropy.
10. The method of claim 1 , wherein the computing device is a mobile device.
11. The method of claim 10 , wherein the plurality of entropy generating sources include one or more of: a magnetometer, an accelerometer, a gyroscope, a microphone, a navigation module, a compass, a camera, a timing clock, a user interface event, a user touch event, a battery level indicator, a power consumption indicator, a wireless access strength, a network speed, a processor temperature, a processor fan speed, an operational environment, a counter, an interruption, a non-deterministic chaos, a touch screen motion, a radio baseband, and, virtual memory statistics.
12. The method of claim 1 , wherein entropy is consumed from the entropy pool when a user interacts with applications installed in the computing device.
13. The method of claim 1 , wherein the computing device has one of the following operating systems: iOS, MacOS, Windows, Linux, Android.
14. The method of claim 13 , wherein the method includes using the encryption algorithms that are customized to be compatible with the operating system of the computing device.
15. The method of claim 1 , wherein the encryption key is backwardly secure, wherein even if the key is compromised, the future output from the entropy pool is unpredictable.