《厦门大学学报（自然科学版）》

Linux统一加密设置(LUKS)是Linux操作系统的标准磁盘加密认证规范并得到广泛应用.由于其算法构成复杂且所需资源较多,如何利用单个现场可编程门阵列(FPGA)的有限资源来实现整个算法并获得高吞吐率是研究工作的重点和难点.为此,研究了一种高能效的LUKS认证算法流水线架构,包括采用4级流水线的安全散列算法(SHA-1)和8级流水线的基于密码的密钥派生函数(PBKDF2)-基于哈希消息验证代码(HMAC)-SHA-1),并使用块随机存取存储器(BRAM)实现了基于S盒和T盒(ST-box)映射表的高级加密标准(AES)-128-电子密码本模式(ECB)算法,以节省FPGA的查找表资源用于上述的流水线架构实现.运行结果显示本设计的密码遍历速度达到了342 s-1,功耗仅为5.27 W,每个密钥的平均计算能量为0.015 J.口令恢复速度超过了工作频率为700 MHz、480核的GTX 480图像处理器(GPU),同时其能耗仅为GPU的1/13.

Linux unified key setup(LUKS)is regarded as the most popular full disk encryption solution implemented in Linux.A high throughput and pipelined architecture of LUKS authentication scheme is implemented on a field-programmable gate array(FPGA)with limited resources,including the four-stage pipeline structure of secure hash algorithm(SHA-1)algorithm,the eight-stage pipeline structure of password-based key derivation function(PBKDF2)-hash-based message authentication code(HMAC)-SHA-1 algorithm and the hardware design of advanced encryption standard(AES)-128-electronic code book(ECB)algorithm based on S-box and T-box(ST-box).Results show that the recovery speed of a password is 342 s-1,power consumption is only 5.27 W,and the average calculation energy per password is 0.015 J.With the authentication system running in the 195 MHz,the password recovery speed is faster than the 480 cores GTX 480 GPU running in 700 MHz,while its power consumption is only 1/13 of the GPU.

引言
1 LUKS算法概述
2 LUKS架构设计与实现
3 实验结果与分析
4 结论

图1 LUKS认证算法的密码验证过程<br/>Fig.1 Password verification process of LUKS authentication algorithm

图1 LUKS认证算法的密码验证过程
Fig.1 Password verification process of LUKS authentication algorithm

表1 流水线结构LUKS中IP核各模块的时延<br/>Tab.1 Amount of delayed for each module of IP core in LUKS pipeline architecture

表1 流水线结构LUKS中IP核各模块的时延
Tab.1 Amount of delayed for each module of IP core in LUKS pipeline architecture

图2 LUKS硬件协处理器架构<br/>Fig.2 Hardware coprocessor architecture of LUKS

图2 LUKS硬件协处理器架构
Fig.2 Hardware coprocessor architecture of LUKS

图3 SHA-1轮操作示意图<br/>Fig.3 Round operation implementation of SHA-1

图3 SHA-1轮操作示意图
Fig.3 Round operation implementation of SHA-1

图4 4级流水线SHA-1<br/>Fig.4 Four-stage pipeline SHA-1

图4 4级流水线SHA-1
Fig.4 Four-stage pipeline SHA-1

表2 4级流水线的SHA-1实现结果比较<br/>Tab.2 Result comparison of four-stage pipelined SHA-1 implementation

表2 4级流水线的SHA-1实现结果比较
Tab.2 Result comparison of four-stage pipelined SHA-1 implementation

图5 算法流程
Fig.5 Flow of algorithm

图6 PBKDF2-HMAC-SHA-1架构<br/>Fig.6 PBKDF2-HMAC-SHA-1 architecture

图6 PBKDF2-HMAC-SHA-1架构
Fig.6 PBKDF2-HMAC-SHA-1 architecture

表3 PBKDF2占用资源及其性能对比<br/>Tab.3 Resource utilization and performance in PBKDF2

表3 PBKDF2占用资源及其性能对比
Tab.3 Resource utilization and performance in PBKDF2

表4 AES算法不同实现之间的数据对比<br/>Tab.4 Comparison between different implementations of the AES algorithm

表4 AES算法不同实现之间的数据对比
Tab.4 Comparison between different implementations of the AES algorithm

图7 流水线结构各模块资源占用率<br/>Fig.7 Resource utilization rate of each module in Pipeline architecture

图7 流水线结构各模块资源占用率
Fig.7 Resource utilization rate of each module in Pipeline architecture

表5 各平台LUKS密码恢复速度和功耗对比(AES-128-ECB)<br/>Tab.5 LUKS password recovery speed and power consumption of each platform(AES-128-ECB)

表5 各平台LUKS密码恢复速度和功耗对比(AES-128-ECB)
Tab.5 LUKS password recovery speed and power consumption of each platform(AES-128-ECB)

[1] MALVONI K,KNEZOVIC' J.Are your passwords safe:energy-efficient bcrypt cracking with low-cost parallel hardware[C]∥WOOT'14 8th Usenix Workshop on Offensive Technologies Proceedings 23rd USENIX Security Symposium.San Diego:USENIX,2014.doi:10.13140/2.1.3267.0081.
[2] LEE H,PAIK Y,JUN J,et al.High-throughput low-area design of AES using constant binary matrix-vector multiplication[J].Microprocessors and Microsystems,2016,47:360-368.
[3] MICHAIL H E,ATHANASIOU G S,KELEFOURAS V I,et al.Area-throughput rrade-offs for SHA-1 and SHA-256 Hash functions' pipelined designs[J].Journal of Circuits,Systems and Computers,2016,25(4):1650032.
[4] Li X,Cao C,Li P,et al.Energy-efficient hardware implementation of LUKS PBKDF2 with AES on FPGA[C]∥Trustcom/BigDataSE/ISPA,Tianjin:IEEE,2016:402-409.
[5] 王文功,于松林,李玉峰,等.基于SHA-1算法的硬件设计及实现[J].电子设计工程,2017,25(4):14-17.
[6] 刘恒,黄凯,修思文,等.多种哈希算法的可重构硬件架构设计[J].计算机工程与科学,2016,38(3):411-417.
[7] 龚向东,王佳,张准,等.基于FPGA的AES算法硬件实现优化及IP核应用[J].电子设计工程,2017,25(12):1-5.
[8] FRUHWIRTH C.LUKS on-disk format specification version 1.2.2.[EB/OL].[2017-8-30].https:∥zh.scribd.com/document/352872322/On-Disk-Format.
[9] BOSSI S,VISCONTI A.What users should know about full disk encryption based on LUKS[C]∥International Conference on Cryptology and Network Security.Marrakesh:Springer,2015:225-237.
[10] STEVENS M,KARPMAN P,PEYRIN T.Freestart collision for full SHA-1[C]∥Annual International Conference on the Theory and Applications of Cryptographic Techniques.Berlin:Springer,2016:459-483.
[11] MICHAIL H,KAKAROUNTAS A,MILIDONIS A,et al.A top-down design methodology for ultrahigh-performance hashing cores[J].IEEE Transactions on Dependable and Secure Computing,2009,6(4):255-268.
[12] MAKKAD R K,SAHU A K.Novel design of fast and compact SHA-1 algorithm for security applications[C]∥IEEE International Conference on Recent Trends in Electronics,Information & Communication Technology(RTEICT).Bangalore:IEEE,2016:921-925.
[13] LEE H,LEE S,KIM J,et al.Parallelizing SHA-1[J].IEICE Electronics Express,2015,12(12):20150371.
[14] VISCONTI A,BOSSI S,RAGAB H,et al.On the weaknesses of PBKDF2[C]∥International Conference on Cryptology and Network Security.Marrakesh:Springer,2015:119-126.
[15] JOHNSON T,ROGGOW D,JONES P H,et al.An FPGA architecture for the recovery of WPA/WPA2 keys[J].Journal of Circuits,Systems and Computers,2015,24(7):1550105.
[16] KUNDI D S,AZIZ A,IKRAM N.A high performance ST-box based unified AES encryption/decryption architecture on FPGA[J].Microprocessors and Microsystems,2016,41:37-46.
[17] KUNDI D S,AZIZ A,KAZMI M.An efficient single unit T-box/T-1-box implementation for 128-bit AES on FPGA[J].Security and Communication Networks,2015,8(9):1725-1731.

备注

引言

1 LUKS算法概述

2 LUKS架构设计与实现

3 实验结果与分析

4 结论

学报简介

备注

引言

1 LUKS算法概述

2 LUKS架构设计与实现

3 实验结果与分析

4 结 论

学报简介

4 结论