TU Darmstadt / ULB / TUbiblio

On the Efficiency of Polynomial Multiplication for Lattice-Based Cryptography on GPUs Using CUDA

Akleylek, Sedat ; Dagdelen, Özgür ; Tok, Zaliha Yüce :
On the Efficiency of Polynomial Multiplication for Lattice-Based Cryptography on GPUs Using CUDA.
In: LNCS , 9540 . Springer
[ Konferenzveröffentlichung] , (2016)

Kurzbeschreibung (Abstract)

Polynomial multiplication is the most time-consuming part of cryptographic schemes whose security is based on ideal lattices. Thus, any efficiency improvement on this building block has great impact on the practicability of lattice-based cryptography. In this work, we investigate several algorithms for polynomial multiplication on a graphical processing unit (GPU), and implement them in both serial and parallel way on the GPU using the compute unified device architecture (CUDA) platform. Moreover, we focus on the quotient ring (Z/pZ)[x]/(xn+1), where p is a prime number and n is a power of 2. We stress that this ring constitutes the most common setting in lattice-based cryptography for efficiency reasons. As an application we integrate the different implementations of polynomial multiplications into a lattice-based signature scheme proposed by Güneysu et al. (CHES 2012) and identify which algorithm is the preferable choice with respect to the ring of degree n.

Typ des Eintrags: Konferenzveröffentlichung ( nicht bekannt)
Erschienen: 2016
Autor(en): Akleylek, Sedat ; Dagdelen, Özgür ; Tok, Zaliha Yüce
Titel: On the Efficiency of Polynomial Multiplication for Lattice-Based Cryptography on GPUs Using CUDA
Sprache: Englisch
Kurzbeschreibung (Abstract):

Polynomial multiplication is the most time-consuming part of cryptographic schemes whose security is based on ideal lattices. Thus, any efficiency improvement on this building block has great impact on the practicability of lattice-based cryptography. In this work, we investigate several algorithms for polynomial multiplication on a graphical processing unit (GPU), and implement them in both serial and parallel way on the GPU using the compute unified device architecture (CUDA) platform. Moreover, we focus on the quotient ring (Z/pZ)[x]/(xn+1), where p is a prime number and n is a power of 2. We stress that this ring constitutes the most common setting in lattice-based cryptography for efficiency reasons. As an application we integrate the different implementations of polynomial multiplications into a lattice-based signature scheme proposed by Güneysu et al. (CHES 2012) and identify which algorithm is the preferable choice with respect to the ring of degree n.

Buchtitel: Cryptography and Information Security in the Balkans
Reihe: LNCS
Band: 9540
Verlag: Springer
Freie Schlagworte: Primitives; P1; Lattice-based cryptography; GPU implementation; CUDA platform; Polynomial multiplication; Fast Fourier transform; cuFFT; NTT; Schönhage-Strassen
Fachbereich(e)/-gebiet(e): 20 Fachbereich Informatik
20 Fachbereich Informatik > Theoretische Informatik - Kryptographie und Computeralgebra
DFG-Sonderforschungsbereiche (inkl. Transregio)
DFG-Sonderforschungsbereiche (inkl. Transregio) > Sonderforschungsbereiche
Profilbereiche
Profilbereiche > Cybersicherheit (CYSEC)
DFG-Sonderforschungsbereiche (inkl. Transregio) > Sonderforschungsbereiche > SFB 1119: CROSSING – Kryptographiebasierte Sicherheitslösungen als Grundlage für Vertrauen in heutigen und zukünftigen IT-Systemen
Veranstaltungsort: Koper, Slovenia
Hinterlegungsdatum: 06 Sep 2018 11:16
Export:

Optionen (nur für Redakteure)

Eintrag anzeigen Eintrag anzeigen