Linux环境中第一行dpa,DPA--差分能量分析初试

DPA BOOK: Power Analysis Attacks - Revealing the Secrets of Smartcards

www.dpabook.org 上有完整的能量轨迹数据和matlab代码

下面仅介绍相关代码：

load('WS3.mat');

% 加载工作区WorkSpace数据，有三个矩阵：

% aes_plaintexts:

%　　明文输入值，大小为1000*16。1000表示1000个明文

%　　由于每次分析8bits，因此128bits明文拆为每列8bits*16列

% traces_noDummy:

%　　实测轨迹，没有插入Dummy保护。大小1000*25000

%　　对1000个明文，每个采样25000个能量泄露点

% traces_withDummy:

%　　实测轨迹，插入Dummy保护。大小1000*25000

samples = 1000;

% 1000个明文即1000条采样轨迹数

analyzed_traces = 'traces_noDummy';

%analyzed_traces = 'traces_withDummy';

byte_to_attack = 1;

%选择攻击的位置，第1个byte。共16个byte

delta = 9441-6729;

more off

%禁用分页输出

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

% Prepare plaintexts

D = aes_plaintexts(1:samples, byte_to_attack);

clear aes_plaintexts byte_to_attack

% Prepare traces

eval(sprintf('traces = %s(1:samples, :);', analyzed_traces));

clear analyzed_traces

% eval(str)：将括号内的字符串str视为可执行语句，并运行

K = uint8(0:255);

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

% TASK 2

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

% Calculate hypothetical intermediate values

V = SubBytes(bitxor(repmat(D, 1, length(K)), repmat(K, samples, 1)) + 1);

% repmat函数见最后

% Calculate hypothetical power consumption

H = HW(V+1);

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

% Calculate correlation

tr_length = size(traces, 2);

R = zeros(length(K), tr_length);

for key_idx = uint16(K)+1

fprintf('Working on key guess = %d\n', K(key_idx));

for k = 1:tr_length

r = corrcoef( [double(H(:,key_idx)) double(traces(:,k))] );

% corrcoef计算两个列向量相关系数，返回协方差矩阵矩阵r

R(key_idx, k) = r(1, 2);

% r(1, 2)为两个向量的相关系数(协方差)

end

end

clear key_idx k r

[rmax,rn]=max(max(R'));

% 返回R中最大值rmax及其行号rn,可得猜测密钥K(rn) ;max函数见最后

% fprintf('Key byte is most likely: %d\n', K(rn));

% plot(R(rn,:));

显示相关系数图像 show_plots函数

exist()

检查变量、脚本、函数、文件夹或类的存在情况

MATLAB的repmat函数

B = repmat(A,r,c)

将A向下(行方向)复制n次，向右(列方向)复制c次。(中间填充的也是矩阵A)

如：

A = 1:4;

B = repmat(A,4,1)

B =

1 2 3 4

1 2 3 4

1 2 3 4

1 2 3 4

MATLAB的max函数