DocumentCode :
2889850
Title :
A parallel unpredictable random number generator
Author :
Marton, K. ; Suciu, A. ; Petricean, D.
Author_Institution :
Tech. Univ. of Cluj-Napoca, Cluj-Napoca, Romania
fYear :
2011
fDate :
23-25 June 2011
Firstpage :
1
Lastpage :
5
Abstract :
Unpredictable random number generators represent a suitable solution when sources of true randomness are not available, too expensive or the nature of the application employing the generator requires a higher throughput and practicality than available TRNGs might provide, but at the same time, the desired level of irreproducibility and unpredictability can not be met by pseudorandom generators. The main goal of the new generator design is to take advantage of the fast randomness generation and different statistical quality of multiple PRNG families and provide an unpredictable combination method, which improves the quality of produced randomness and mitigates certain security problems the individual generators are exposed to if used on their own. Furthermore, the proposed generator leverages the processing power of multi-core and SMP systems by enabling the parallel execution of multiple generators selected by the user´s unpredictable mouse movements. The result is a stream of unpredictably combined pseudorandom number sequences which shows high randomness quality - evaluated by the well known NIST statistical test suite, improved by the unpredictability and nondeterminism characteristic to human-computer interaction and doubled by a high generation speed provided by the parallel generation strategy. These characteristics enable the generator to be considered suitable for producing high quality randomness with the promise of cryptographic security.
Keywords :
cryptography; human computer interaction; multiprocessing systems; random number generation; statistical testing; NIST statistical test suite; PRNG families; SMP systems; cryptographic security; human-computer interaction; multicore systems; parallel generation strategy; parallel unpredictable random number generator; statistical quality; Cryptography; Generators; Lattices; Mice; NIST; Throughput; human-computer interaction; random number generator; statistical testing; unpredictability;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Roedunet International Conference (RoEduNet), 2011 10th
Conference_Location :
Iasi
ISSN :
2068-1038
Print_ISBN :
978-1-4577-1233-3
Type :
conf
DOI :
10.1109/RoEduNet.2011.5993701
Filename :
5993701
Link To Document :
بازگشت