Separate encrypt/decrypt

[ajitam]

Hi, so I continue my work using you script and I hit a problem. I'm not 100% this is an issue, because the logic changed and I'm still trying to wrap my head around the code. So here we go:

I prepared a simple script which:

• encrypts 2bytes (0xAA 0xEE)

• uses 10 iterations

  #include <Cape.h>
  Cape cape("oEpHgSPFvZ6TzrzkaKMX", 10);
  
  #define DO_ENCRYPT 1
  #define DO_DECRYPT 1
  
  void setup() {
    Serial.begin(115200);
  }
  
  void loop() {
  
  #if DO_ENCRYPT
    char in[] = {0xAA, 0xEE};
  
    cape.encrypt(in, 2);
    byte enPacket_1 = cape.result[0];
    byte enPacket_2 = cape.result[1];
    byte enPacket_3 = cape.result[2];
  
    Serial.print("0x");
    Serial.print(enPacket_1, HEX);
    Serial.print(" | ");
    Serial.print(enPacket_2, HEX);
    Serial.print(" | ");
    Serial.println(enPacket_3, HEX);
  
    delay(1000);
  #endif
  
  #if DO_DECRYPT
    char out[] = {enPacket_1, enPacket_2, enPacket_3};
    //char out[] = {0xC9, 0xE2, 0x3B};
  
    cape.decrypt(out, 2);
    byte dePacket_1 = cape.result[0];
    byte dePacket_2 = cape.result[1];
  
    Serial.print("0x");
    Serial.print(dePacket_1, HEX);
    Serial.print(" | ");
    Serial.println(dePacket_2, HEX);
  
    delay(1000);
  #endif
  
  }

Steps

1. If I set DO_ENCRYPT 1 and DO_DECRYPT 1 I get this

   0xC1 | EA | 33
   0xAA | EE
   0x81 | AA | 73
   0xAA | EE
   0x5B | 70 | A9
   0xAA | EE
   0x17 | 3C | E5
   0xAA | EE
   0xDF | F4 | 2D
   0xAA | EE
   0xEF | C4 | 1D
   0xAA | EE
   0xAB | 80 | 59
   0xAA | EE
   

   which is correct.

2. Then I set set DO_ENCRYPT 1 and DO_DECRYPT 0 and I get this

   0xC1 | EA | 33
   0x9 | 22 | FB
   0x15 | 3E | E7
   0xAB | 80 | 59
   0xEB | C0 | 19
   0x23 | 8 | D1
   0x3B | 10 | C9
   0x7B | 50 | 89
   

3. I then:

   • select one of the (encrypted) values (lets say 0xAB 80 59) and I replace char out[] with this values like so char out[] = {0xAB, 0x80, 0x59};
   • set DO_ENCRYPT 0 and DO_DECRYPT 1
     and I get this

   0xF3 | B7
   0xB6 | F2
   0xB6 | F2
   0xB6 | F2
   0xB6 | F2
   

Problem

Any ideas why data is not decrypted to the same value. As far I understand 3th byte is initialization byte which just "mangles" data.

[gioblu]

@ajitam Hi! Thank you again for your support and betatest.
You tested an unreleased not stable version of Cape.

I this week end and today worked to release a new stable version with a different and more secure encryption method. After various tests I decided to get rid of iterations, to save computation time (because of multiple function call). I ended up with a 3 step non-iteration-tunable encryption vastly based on private key and data hashing.

I also used your example to doublecheck its correct functionality:

#include <Cape.h>
Cape cape("oEpHgSPFvZ6TzrzkaKMX", 20);
#define DO_ENCRYPT 1
#define DO_DECRYPT 1

void setup() {
  Serial.begin(115200);
}

void loop() {
#if DO_ENCRYPT
  char in[] = {10, 10};
  cape.encrypt(in, 2);
  byte enPacket_1 = cape.result[0];
  byte enPacket_2 = cape.result[1];
  byte enPacket_3 = cape.result[2];

  Serial.print(enPacket_1);
  Serial.print(" | ");
  Serial.print(enPacket_2);
  Serial.print(" | ");
  Serial.println(enPacket_3);

  delay(1000);
#endif

#if DO_DECRYPT
  char out[] = {enPacket_1, enPacket_2, enPacket_3};

  // From static crypted string back to 10, 10
  //char out[] = {108, 65, 13};

  cape.decrypt(out, 3);
  byte dePacket_1 = cape.result[0];
  byte dePacket_2 = cape.result[1];

  Serial.print(dePacket_1);
  Serial.print(" | ");
  Serial.println(dePacket_2);
  delay(1000);
#endif
}

Changelog:

• In Cape definition, you have to pass (as explained in the updated readme) the key and its length.
• You have to pass to decrypt the real length of the string (counting also the initialization vector).

I am sorry for the temporary instability of the code.

[gioblu]

@ajitam https://github.com/gioblu/Cape/releases/tag/1.2

[ajitam]

Is this new encryption based on some other method? (previous one was similar to RC4)

[gioblu]

Hi @ajitam. I would not indicate any method in particular as "inspiration" for this. Other users complained because RC4 is holed like Emmenthal, so I decided to work on a new, dedicated method based on standard and proved effective methods to encrypt a string.

[gioblu]

@ajitam I tried to be as clear as possible in the code's comments.
Let me know what do you think about it.
Distribution seems really better than before.

[ajitam]

I'm not an expert in cryptology so frankly - I don't know which method is better and what are there flaws.
Thing is that currently I'm searching for the encryption to encrypt 2 or 3 bytes and I have 4 bytes of storage.
Buy my setup is a bit special:

• encrypted data is always different
• encryption key is always different

(closing this issue - sorry I didn't do it earlier)