Cristina-Loredana Duta

Accelerating Encryption Algorithms Using Parallelism

The importance of protecting the information has increased rapidly during the last decades and as a consequence so did the need for cryptographic algorithms. We want to increase the speed of these methods that protect our data as much as we can. In this paper , we used parallelism for encryption algorithms and I/O operations to emphasize the benefit of it, by implementing and improving two cryptographic modes such as Cipher Block Chaining (CBC ) and Interleaved CBC (ICBC ) for Advanced Encryption Algorithm (AES ). The result of this paper is the remarkable speedup in encryption and decryption time of cryptographic systems when using parallel paradigms (OpenCL, Cuda, OpenMP) and the comparison of the serial versus parallel implementation.

Real time implementation of MELP speech compression algorithm using Blackfin processors

A large part of the latest research in speech coding algorithms is motivated by the need of obtaining secure military communications, to allow effective operation in a hostile environment. Since the bandwidth of the communication channel is a sensitive problem in military applications, low bit-rate speech compression methods are mostly used. Several speech processing applications such as Mixed Excitation Linear Prediction are characterized by very strict requirements in power consumption, size, and voltage supply. These requirements are difficult to fulfill, given the complexity and number of functions to be implemented, together with the real time requirement and large dynamic range of the input signals. To meet these constraints, careful optimization should be done at all levels, ranging from algorithmic level, through system and circuit architecture, to layout and design of the cell library. The key points of this optimization are among others, the choice of the algorithms, the modification of the algorithms to reduce computational complexity, the choice of a fixed-point arithmetic unit, the minimization of the number of bits required at every node of the algorithm, and a careful match between algorithms and architecture. This paper concentrates on low bit rate speech coding technology, mainly in MELP and solved the problem of optimizing the program of MELP on Digital Signal Processor platform. The algorithm was ported onto a fixed point DSP, Blackfin 537, and stage by stage optimization was performed to meet the real time requirements. The main functions involved were analysis, parameter encoding, parameter decoding and synthesis. The fixed point source code at the MELP front end was also thoroughly optimized at the C Level. Memory optimization techniques such as data placement and caching were also used to reduce the processing time. The results we obtained show that real-time implementations of a speech vocoder based on the MELP standard for low bit rate communications (2400 bps) can be successful on DSP platforms.

Real-time DSP Implementations of Voice Encryption Algorithms.

In the last decades, digital communications and network technologies have been growing rapidly, which makes secure speech communication an important issue. Regardless of the communication purposes, military, business or personal, people want a high level of security during their conversations. In this context, many voice encryption methods have been developed, which are based on cryptographic algorithms. One of the major issues regarding these algorithms is to identify those that can ensure high throughput when dealing with reduced bandwidth of the communication channel. A solution is to use resource constrained embedded systems because they are designed such that they consume little system resources, providing at the same time very good performances. To fulfil all the strict requirements, hardware and software optimizations should be performed by taking into consideration the complexity of the chosen algorithm, the mapping between the selected architecture and the cryptographic algorithm, the selected arithmetic unit (floating point or fixed point) and so on. The purpose of this paper is to compare and evaluate based on several criteria the Digital Signal Processor (DSP) implementations of three voice encryption algorithms in real time. The algorithms can be divided into two categories: asymmetric ciphers (NTRU and RSA) and symmetric ciphers (AES). The parameters taken into consideration for comparison between these ciphers are: encryption, decryption and delay time, complexity, packet lost and security level. All the previously mentioned algorithms were implemented on Blackfin and TMS320C6x processors. Making hardware and software level optimizations, we were able to reduce encryption/decryption/delay time, as well as to reduce the energy consumed. The purpose of this paper is to determine which is the best system hardware (DSP platform) and which encryption algorithm is feasible, safe and best suited for real-time voice encryption.

Performance Comparison of Voice Encryption Algorithms Implemented on Blackfin Platform

Nowadays, secure speech communications have become a fundamental issue, due to the rapid development of digital communications and networking technologies. People want high security level during their communications, especially for military and business purpose. For instance, in order to operate properly and effective in a hostile environment is mandatory to have a secure military communication. The methods that are currently available for voice encryption include the use of cryptographic algorithms (which ensure secure data transmission and reception). However, taking into consideration the problem of the bandwidth of the communication channel (which appears frequently in military applications), it is important to identify encryption algorithms that ensure high throughput and low bit-rate speech compression methods. Various speech encryption methods have strict requirements such as power consumption, size, voltage supply, which are difficult to fulfil. A solution to meet these constraints is to perform optimization at all levels, starting from the algorithm design, continuing with the system and circuit structure and reaching also the design of the cell library. More exactly, these optimizations depend on the modification of the encryption algorithm selected for the applications (in order to reduce its computational complexity), the algorithm chosen, the arithmetic unit selected (fixed-point or floating-point), the mapping between the selected architecture and the cryptographic algorithm. This paper describes a comparison regarding the performance of various speech encryption algorithms implemented on Digital Signal Processor (DSP) platform. These algorithms are from different categories, such as: stream ciphers (Grain v1, Trivium, Mickey 2.0 and SOSEMANUK), block ciphers (Advanced Encryption Standard - AES and Data Encryption Standard - DES) and dedicated algorithms for voice encryption (Robust Secure Coder, scrambling encryption and encryption algorithm based on chaotic map and Blowfish). All the previously mentioned algorithms where implemented on Blackfin 537 (a fixed point DSP) and careful optimizations were performed to fulfil real time requirements. The goal of this paper was to evaluate all these algorithms to determine which one is the best suited for applications that require secure real time communications.

Performance Analysis of Real Time Implementations of Voice Encryption Algorithms using Blackfin Processors

A large part of the latest research in speech coding and speech encryption algorithms is motivated by the need of obtaining secure military communications, to allow effective operation in a hostile environment. Since the bandwidth of the communication channel is a sensitive problem in military applications, low bitrate speech compression methods and high throughput encryption algorithms are mostly used. Several speech encryption methods are characterized by very strict requirements in power consumption, size, and voltage supply. These requirements are difficult to fulfill, given the complexity and number of functions to be implemented, together with the real time requirement and large dynamic range of the input signals. To meet these constraints, careful optimization should be done at all levels, ranging from algorithmic level, through system and circuit architecture, to layout and design of the cell library. The key points of this optimization are among others, the choice of the algorithms, the modification of the algorithms to reduce computational complexity, the choice of a fixed-point arithmetic unit, the minimization of the number of bits required at every node of the algorithm, and a careful match between algorithms and architecture. This paper describes the performance analysis on Digital Signal Processor (DSP) platform of some of the recently proposed voice encryption algorithms, as well as the performance of stream ciphers such as Grain v1, Trivium and Mickey 2.0 (which are suited for real time voice encryption). The algorithms were ported onto a fixed point DSP, Blackfin 537, and stage by stage optimization was performed to meet the real time requirements. Memory optimization techniques such as data placement and caching were also used to reduce the processing time. The goal was to determine which of the evaluated encryption algorithms is best suited for real time secure communications.

Framework for evaluation and comparison of integer factorization algorithms

Mathematicians have always been fascinated by the subject of finding the prime numbers of large composite numbers. They have studied various methods and have focused on developing several theories. Nowadays, this problem is of practical importance because the security of the public key cryptosystems depends on the difficulty of factoring public keys. The integer factorization algorithms have rapidly extended and improved such that today it is easy to factor a 100-decimal digit number and feasible to factor 155-decimal digit number. But the problem of factoring large integers still remains difficult because none of the algorithms run in polynomial time and numbers with more than 150 decimal digits are hard to factor. In this paper, we describe the implementation and performance of several integer factorization algorithms, in order to determine which is more efficient: trial division, Eulers, Fermats, Pollards p-1, Williams p+1, Pollard Rho, Elliptic Curve Method, Shanks square forms, Dixons, Continued Fraction Factorization, Quadratic Sieve, Number Field Sieve, Special Number Field Sieve and General Number Field Sieve. We implemented the algorithms and calculated their approximated computational runtime for input numbers of various sizes. Then, we mathematically analyzed their time complexities and then, to determine their performance, we compared the results of each algorithm with each other. For the implementation of the algorithms we have built an evaluation framework that contains the previously mentioned factorization algorithms and that allows the user to load data of different input size.

Framework for Evaluation and Comparison of Primality Testing Algorithms

Nowadays, public-key cryptography is widely used to ensure data protection (privacy, integrity, authentication). Public-key algorithms contain, as an essential part of their construction, the usage of prime numbers and of large composite numbers, which have prime factors. Because of this, the security of these algorithms such as RSA (Rivest-Shamir-Adelman) depend on the difficulty to factorize the public keys. For many centuries, mathematicians have been fascinated by the subject of prime numbers, have studied their proprieties and have focused on developing several theories. Many primality tests, deterministic and probabilistic, were proposed to determine whether a number is prime or not. In this paper, we describe the implementation and performance of several primality tests, in order to determine which is more efficient: Fermats test, Miller-Rabin test, Solovay-Strassen test, Agrawal-Kayal-Saxena (AKS) test, Lucas-Lehmer test, Baillie-PWS test, Pepins test, Lucas-Lehmer-Riesel (LLR) test, Proths theorem, Quadratic Frobenius test, Adleman-Pomerance-Rumley (APR) test, Lucas test or Pocklington test. The tests were implemented using .NET framework (C#) and the performance was determined based on input numbers of various sizes and on the type of the primality test (deterministic/probabilistic). These algorithms were also analyzed taking into account their ability to correctly determine the primality of a given number.