INCREASING SPEED OF ELLIPTIC CURVE ELGAMAL WITH PARALLELIZATION

• Main Author: Suwanda, Hokki
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/73900
• Institution: Institut Teknologi Bandung

To attain higher security level, public key cryptography like RSA uses bigger key size. However, elliptic curve cryptography like Elliptic Curve ElGamal uses much smaller key size compared to RSA to attain the same level of security. Even so, the key size remains big. It could cause encrypting and decrypting messages to take more time. One way to increase the speed of data processing is parallelization. Parallelization can be done in CPU. There are several variations of parallelization in CPU, they are distributed-memory programs and shared- memory programs. Another variation could be made by mixing both distributed- memory and shared-memory programs, called hybrid parallel programs. The solution implemented will be based on the three variations. Each program outputs the encryption and decryption duration in seconds. The duration will then be used to calculate throughput, the number of characters processed each second. Throughput is further used to calculate speedup. The experiment is to determine whether the solution is successful in increasing speed. From the result, it was found that all solutions are capable of increasing algorithm speed, with shared- memory parallel program being the best, followed by distributed-memory program and hybrid parallel program.