Search code examples
cryptographyrfcrfc-4226

Is this RFC 4226 wrong?

The test values of the RFC specify:

Appendix D - HOTP Algorithm: Test Values

   The following test data uses the ASCII string
   "12345678901234567890" for the secret:

   Secret = 0x3132333435363738393031323334353637383930

   Table 1 details for each count, the intermediate HMAC value.

   Count    Hexadecimal HMAC-SHA-1(secret, count)
   0        cc93cf18508d94934c64b65d8ba7667fb7cde4b0
   1        75a48a19d4cbe100644e8ac1397eea747a2d33ab

So if I try to get the HMAC for 0 in ruby I get:

[20] pry(AuthyOTP)> secret_key = "12345678901234567890"
=> "12345678901234567890"
[22] pry(AuthyOTP)> OpenSSL::HMAC.hexdigest(digest, secret_key, "0")
=> "32a67f374525d32d0ce13e3db42b5b4a3f370cce"

I was expected to get cc93cf18508d94934c64b65d8ba7667fb7cde4b0

So I wrote an implementation in java, and I get the same:

Calculation OTP for movingFactor = 0
    2. Calculate Hash = 
      32a67f374525d32d0ce13e3db42b5b4a3f370cce

So what is the hexadecimal SHA1-HMAC of "0" when the secret is "12345678901234567890" ?

Solution

  • RFC4226 is correct.

    You are confusing a character strings with bytes. You are not suppose to compute the hmac-sha1 of '0', you are suppose to compute the hmac-sha1 of an 8 byte integer that starts out at 0. In java, that would be the hmac-sha1 of byte [] counter = {0, 0, 0, 0, 0, 0, 0, 0};