假如Alice和Bob分别利用公钥e_1 ,e_2对明文M进行加密:
C_1=M^(e_1 ) (mod n)
C_2=M^(e_2 ) (mod n)
则可以分别利用私钥d_1 ,d_2进行解密:
M=C_1^(d_1 ) (mod n)
M=C_2^(d_2 ) (mod n)
假如攻击者监听获得了密文C1,C2,所有公钥都是公开的,那么攻击者在已知C1,C2,e1,e2,n的情况下,可以绕过d1,d2直接恢复出明文M。利用信息安全数学基础证明如下:
首先假设这些公钥共模且互素,即有gcd(e_1,e_2 )=1,则
e_1x+e_2y=1
其中x,y皆为整数,一正一负;利用欧几里得扩展算法可以求得一组解(x,y),假设x为正,y为负,利用模运算的性质,因为
成功恢复明文M.
以下我以最近做过的一道CTFcrypt题目为例,演示RSA公共模数攻击,题目如下图,模数n长度为4096bit,
附其源码veryhardRSA.py文件,公钥分别为e1=17,e2=65537
#!/usr/bin/env python
import random
N = 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
def pad_even(x):
return ('', '0')[len(x)%2] + x
e1 = 17
e2 = 65537
fi = open('flag.txt','rb')
fo1 = open('flag.enc1','wb')
fo2 = open('flag.enc2','wb')
data = fi.read()
fi.close()
while (len(data)<512-11):
data = chr(random.randint(0,255))+data
data_num = int(data.encode('hex'),16)
encrypt1 = pow(data_num,e1,N)
encrypt2 = pow(data_num,e2,N)
fo1.write(pad_even(format(encrypt1,'x')).decode('hex'))
fo2.write(pad_even(format(encrypt2,'x')).decode('hex'))
fo1.close()
fo2.close()
写个脚本,利用前述分析过程:
# encoding: utf-8
import sys
sys.setrecursionlimit(1000000) #设置为一百万
N = 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
def pad_even(x):
return ('', '0')[len(x)%2] + x
#欧几里得扩展算法
def extgcd(a, b):
if a == 0:
return (b, 0, 1)
else:
g, y, x = extgcd(b % a, a)
return (g, x - (b // a) * y, y)
#求模逆
def modinv(a, m):
g, x, y = extgcd(a, m)
if g != 1:
raise Exception('modular inverse does exist!')
else:
return x % m
def main():
e_1 = 17
e_2 = 65537
fo1 = open('flag.enc1', 'rb')
fo2 = open('flag.enc2', 'rb')
fo3 = open('flag_dec.txt', 'wb')
data1 = fo1.read()
data2 = fo2.read()
fo1.close
fo2.close
c1 = int(data1.encode('hex'), 16)
c2 = int(data2.encode('hex'), 16)
g, x, y = extgcd(e_1, e_2)
#求模反元素
if x < 0:
x = - x
c1 = modinv(c1, N)
elif y < 0:
y = - y
c2 = modinv(c2, N)
if g == 1:
m = (pow(c1, x, N) * pow(c2, y, N)) % N
fo3.write(pad_even(format(m,'x')).decode('hex'))
fo3.close
if __name__ == '__main__':
main()
查看结果:
成功获取到PCTF{M4st3r_oF_Number_Th3ory},解密成功。
可见,在攻击者已知C1,C2,e1,e2,N时,可以在不知道私钥d1,d2的情况下恢复出明文M,如果不同的用户使用相同的模数。