The entropy of an algorithm measures how strong the algorithm is against cryptographic attacks. How do we mathematically measure the entropy of text cipher?
Hi Thenmozhi,
Greetings, there are different models for you your mean. At first, you need to choose an suitable model. I suggest to you that use the standard model established in the following paper:
1.http://www.ijresm.com/Vol_1_2018/Vol1_Iss10_October18/IJRESM_V1_I10_43.pdf
: Waghmare, Wrushali, S. I. Nipanikar, and P. V. Mulmule. "Holo-Entropy and Advanced Encryption Standard for Wavelet-Based Image Steganography." International Journal of Research in Engineering, Science and Management 1.10 (2018).
Regards, Saeed
Hi Thenmozhi,
Entropy tests are really important in cryptography. A perfect Pseudorandom Random Permutation (PRP) have very high entropy (close to 8 bits) , however high entropy is necessary but not sufficeient criteria for a good cipher, or good random number generator.
For example this very small python code using LCG random number generator which is cryptographically insecure, however producing very high entropy outptut.
----------------------------------------------------------------
import struct
v=7;
f=open('rnd2.bin','wb');
for i in range(500000):
r=v=v**5%(2**64-59)
f.write(struct.pack('B',r>>24&255))
----------------------------------------------------------------
OUTPUT test:
test$ ent rnd2.bin
Entropy = 7.999699 bits per byte.
Optimum compression would reduce the size
of this 500000 byte file by 0 percent.
----------------------------------------------------------------
I uploaded a simple entropy test and you can see that this algorithm created random numbers with very high entropy. "7.999699" . (I used the "ent" linux command to measure entropy). However this generator is not secure at all, and definitely not a Pseudorandom Random Permutation.
This is a good example that you can not measure the strength of an algorithm
(e.g: cipher, random number generator) relying only on entropy. There are many other aspects that you have to consider.
I hope this answer your question.
Regards,
Norbert
Norbert Tihanyi Thank u sir for the input provided. Avalanche effect and Entropy i have considered as of now. What are the other aspects that are to be considered?
- Badges
- Science topic
- Similar topics
- Human-Computer Interaction
- Collaboration