The effective finite-field inversion (FF-inversion), a computationally intensive technique that significantly affects overall system performance, is necessary for Elliptic Curve Cryptography (ECC) and related cryptographic applications. Among inversion methods, the Itoh-Tsujii algorithm (ITA) is especially attractive for FPGA implementations because it replaces long Euclidean sequences with structured squaring and multiplications. This paper introduces High Speed-ITA (HS-ITA), an optimized ITA variant that uses 16n (hex) power blocks and precomputed addition chains to reduce sequential exponentiation depth and enable parallel evaluation on FPGAs. The use of specialized hex-power blocks, combined with optimized reuse of intermediate powers, allows the design to complete inversion in GF(2233) within 24 cycles, without significantly increasing hardware resources.Virtex-7, Virtex-4, and Virtex-5 devices are used for implementation, the design attains improved area-delay product and throughput compared with prior ITA and EEA FPGA implementations. The architecture trades a small increase in LUT usage for significantly lower latency, making it appropriate for resource-constrained high-performance cryptographic systems.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Fast Modular Inversion in GF(2m) on FPGA Using Optimized Exponentiation

  • J. Adline Vidhya,
  • V. R. Venkatasubramani,
  • S. Rajaram,
  • V. Vinoth Thyagarajan

摘要

The effective finite-field inversion (FF-inversion), a computationally intensive technique that significantly affects overall system performance, is necessary for Elliptic Curve Cryptography (ECC) and related cryptographic applications. Among inversion methods, the Itoh-Tsujii algorithm (ITA) is especially attractive for FPGA implementations because it replaces long Euclidean sequences with structured squaring and multiplications. This paper introduces High Speed-ITA (HS-ITA), an optimized ITA variant that uses 16n (hex) power blocks and precomputed addition chains to reduce sequential exponentiation depth and enable parallel evaluation on FPGAs. The use of specialized hex-power blocks, combined with optimized reuse of intermediate powers, allows the design to complete inversion in GF(2233) within 24 cycles, without significantly increasing hardware resources.Virtex-7, Virtex-4, and Virtex-5 devices are used for implementation, the design attains improved area-delay product and throughput compared with prior ITA and EEA FPGA implementations. The architecture trades a small increase in LUT usage for significantly lower latency, making it appropriate for resource-constrained high-performance cryptographic systems.