Resource-Shared Crypto-Coprocessor of AES Enc/Dec With SHA-3
Resource-Shared Crypto-Coprocessor of AES Enc/Dec With SHA-3
复制标题
具有 SHA-3 的 AES Enc/Dec 资源共享加密协处理器
DOI：
10.1109/tcsi.2020.2997916
复制
发表时间：
2020-12-01
影响因子：
5.1
通讯作者：
Liu, Weiqiang
中科院分区：
文献类型：
--
作者：
Kundi, Dur-e-Shahwar;Khalid, Ayesha;Liu, Weiqiang
密码协处理器是现代片上系统不可或缺的一部分。这种设计的灵活性具有双重目的，即，它使得能够加速不同的基本密码原语（加密/认证/伪随机数生成（PRNG）），并且还通过资源共享导致设计紧凑。在此背景下，提出了一种新的资源共享密码协处理器，命名为AE$HA-3，它结合了两个国家标准与技术研究所（NIST）的标准化算法，即，高级加密标准（AES）和安全哈希算法-3（SHA-3）。由于算法的不同，到目前为止还没有资源共享的实现，使AES密钥调度/ enc/dec和SHA-3已经提出。AE$HA-3利用资源共享来减少面积，即，用于AES enc/dec的查找表（I-Tables）的集成;用于SHA-3的六输入方程（SixIE）的逻辑优化;用于在AES和SHA-3变换中执行密钥白化的统一XOR部分。此外，AES密钥调度使用相同的资源共享硬件执行。与迄今为止最小的独立实现相比，Xilinx Virtex FPGA系列上的AE$HA-3在每片吞吐量（TPS）方面具有最高的硬件效率，沿着49.37%的面积消耗减少。
Cryptographic co-processors are integral to the modern System-on-Chips. Flexibility in such designs serves dual purpose, i.e., it enables acceleration of different essential cryptographic primitives (Encryption/Authentication/Pseudo Random Number Generation (PRNG)) and also results in design compaction via resource sharing. In this context, a novel resource-shared crypto-coprocessor, named AE$HA-3 is presented, which combines two National Institute of Standards and Technology (NIST) standardized algorithms, i.e., Advance Encryption Standard (AES) and Secure Hash Algorithm-3 (SHA-3). Due to algorithmic dissimilarities, so far no resource-shared implementation enabling AES key scheduling/ enc/dec and SHA-3 has been presented. AE$HA-3 exploits resource-sharing for area reduction, i.e., integration of Look-Up-Tables (I-Tables) for AES enc/dec; logical optimization of Six Input Equation (SixIE) for SHA-3; a Unified XOR Section to carry out both key whitening in AES and SHA-3 transformations. Furthermore, the AES key scheduling was performed using the same resource-shared hardware. The proposed AE$HA-3 on Xilinx Virtex FPGA family results in highest hardware efficiency in terms of Throughput per Slice (TPS), along with a 49.37% area consumption reduction, when compared against the smallest stand-alone implementations presented to date.