Pseudorandom number generators using multiple sources of entropy

Date

2010-01-21T22:27:25Z

Authors

Srivastava, Gautam

Journal Title

Journal ISSN

Volume Title

Publisher

Abstract

Randomness is an important part of computer science. A large group of work, both in theoretical and practical computer science, is dedicated to the study of whether true 'randomness' is necessary for a variety of applications and protocols to work. One of the main uses for randomness is in the generation of keys, used as a security measure for many cryptographic protocols. The main measure of randomness is achieved by looking at entropy, a measure of the disorder of a system. Nature is able to provide us with many sources that are high in entropy. However, many cryptographic protocols need sources of randomness that are stronger (higher in entropy) than what is present naturally to ensure security. Therefore, a gap exists between what is available in Nature, and what is necessary for provable security. This paper looks to bridge this gap. Research in pseudorandom number generation has gone on for decades. However, many of the past constructions were lacking in either documentation or provable security of their methods. The need for a pseudorandom number generator (PRG) with provable security and strong documentation is evident. A new construction of a PRG is introduced. The new construction, labeled XRNG, looks to encompass recent research in the field of extractors along with previously known research in the field of pseudorandom number generation. Extractors, as the name suggests, looks to extract close to random information from high entropy sources.

Description

Keywords

random number generators, data encryption

Citation