A Distributed Multiprecision Cryptographic Library Design

Abstract

Cryptographic schemes require specialized software libraries to work with large numbers on fixed-precision processors. The concept is known as multiple-precision computation. In this thesis, we aim to review the multiple-precision algorithms with the contemporary modifications. With this motivation, we develop a new multiprecision library named CRYMPIX and we carefully benchmark CRYMPIX in comparison with the fastest alternatives. We also develop a distributed wrapper for computationally expensive functions. Hence, we provide an abstraction method for the higher level cryptographic implementations by allowing them run in a distributed environment without containing any specialized code for distribution.