...
首页> 外文期刊>IEEE Transactions on Computers >Arithmetical Improvement of the Round-Off for Cryptosystems in High-Dimensional Lattices
【24h】

Arithmetical Improvement of the Round-Off for Cryptosystems in High-Dimensional Lattices

机译:高维格上密码系统四舍五入的算术改进

获取原文
获取原文并翻译 | 示例
           

摘要

With Lattice-based cryptography (LBC), ciphertexts are represented as points near a lattice, and Babai's round-off algorithm allows to decrypt them when one knows the secret-key. Recently, an accelerated variant of the round-off, based on Residue Number Systems (RNSs), has been proposed. Herein, we combine this technique with the use of lattices of Optimal Hermite Normal Form (OHNF) and propose further refinements, so as to reduce the decryption complexity. This approach lends itself largely to data-level parallelism, allowing for low latency decryption operations on multi-core CPUs with Single Instruction Multiple Data (SIMD) extensions, and achieves high-throughput on GPUs. Finally, we are able to perform decryptions up to 20 times faster than the most efficient implementation in related art, which exploits the Mixed-Radix System (MRS), in an Intel i7 6700K CPU, and we are able to decrypt up to 11,832 messages/s in a Titan X GPU.
机译:使用基于格的加密(LBC),密文表示为点阵附近的点,并且Babai的舍入算法允许在知道密钥的情况下对它们进行解密。近来,已经提出了基于残数系统(RNS)的四舍五入的加速变体。在本文中,我们将这种技术与最佳厄米正常范式(OHNF)的格网结合使用,并提出了进一步的改进方案,以降低解密的复杂性。这种方法在很大程度上适合于数据级并行性,从而允许在具有单指令多数据(SIMD)扩展的多核CPU上进行低延迟的解密操作,并在GPU上实现高吞吐量。最终,我们能够以比现有技术中采用混合基数系统(MRS)的最有效实现更快的速度执行20倍的解密,而该技术利用Intel i7 6700K CPU,并且我们最多可以解密11,832条消息/ s在Titan X GPU中。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号