Cryptology ePrint Archive: Report 2018/595

Randomness analysis for multiple-recursive matrix generator

Subhrajyoti Deb and Bubu Bhuyan, and Sartaj Ul Hasan

Abstract: Randomness testing of binary sequences generated by any keystream generator is of paramount importance to both designer and attacker. Here we consider a word-oriented keystream generator known as multiple-recursive matrix generator, which was introduced by Niederreiter (1993). Using NIST statistical test suite as well as DieHarder statistical package, we analyze randomness properties of binary sequences generated by multiple-recursive matrix generator and show that these sequences are not really adequate for any cryptographic applications. We also propose nonlinearly filtered multiple-recursive matrix generator based a special nonlinear function and establish that sequences generated by such a nonlinearly filtered multiple-recursive matrix generator provide good randomness results. Moreover, we compare our randomness test results with some of the recent lightweight stream ciphers like Lizard, Fruit, and Plantlet.

Category / Keywords: ecret-key cryptography / Linear feedback shift register; stream cipher; randomness; word-oriented feedback shift register; MRMG; NIST STS; DieHarder

Date: received 9 Jun 2018, last revised 17 Jul 2018, withdrawn 17 Jul 2018

Contact author: sartaj hasan at iitjammu ac in

Available format(s): (-- withdrawn --)

Version: 20180717:144303 (All versions of this report)

Short URL: ia.cr/2018/595


[ Cryptology ePrint archive ]