DYNALYSIS NIST TEST RESULTS OVERVIEW

DynAlysis has tested the distribution complexity of JumPedal’s one-time pad ciphers and their proficiency in conforming to rigorous sequencing evaluations.

DynAlysis has reviewed JumPedal’s ability to utilize software driven deterministic processes to transform 16-byte sequences into randomized one-time pad ciphers. We have tested these ciphers for their ability to pass NIST SP800-22 Revision 1a Statistical Test Suite for Randomness.

Even though there is no absolute test for randomness, the NIST tests are generally accepted as the benchmark for randomness testing. If any of the 15 tests fail, the analyzed sequence fails the requirements of the one-time pad evaluations.

The test results include pass or fall on all 15 tests, along with graphs of zero-one run lengths along with their byte distributions.

Test results include:

  • NIST test specifications and mathematical processes
  • Input file name
  • Bit length recorded in bytes, there are 8 bites per byte
  • Input file type
  • One-time pad 16-byte control sequence
  • Number of sub-blocks, and sub-block byte lengths
  • Input file’s information content for each sub-block pre-encoding
  • One-time pad cipher for each sub-block
  • One-time pad and XOR encoding result for each sub-block

EXPLANATION OF INPUT FILE SUB-BLOCK DIVISSIONS

A variety of 360 files, that range in size from less than 500-KB to more than 3-megabytes, have been evaluated.

Input files have been arranged into sub-block groupings to produce test evaluations that conform to the encoded block length standards of the transmission packets on JumPedal’s CipherCorps network.

Sub-block arrangements of 131,072-bytes have been created from input files that are longer than 131,072 bytes, and 16,384-bytes when the input is less than 131,072-bytes. The last block of every sub-block group has been filled with noise to produce uniform length sub-blocks that are generally rounded to the 16,384-byte packet size.

As an example, a 1.24 megabyte size file will be sub-divided into nine sub-blocks of 131,072-bytes, and ten sub-blocks of 16,384-bytes. The last of the ten will be filled to the 16,384-byte length with noise. Each of the one-time pad ciphers that are required to cipher the megabyte of information have been generated from a single 16-byte control sequence. The rule is one 16-byte control sequence for the encoding of each input file regardless of how many sub-blocks divisions are constructed from the file.

DYNALYSIS OBSERVER RESULTS

DynAlysis has observed that JumPedal’s one-time pad creator:

  • Transforms a single input control sequence or digital seed of 16-bytes into tens-to-hundreds of randomized file sub-block ciphers.
  • Has performed without error on test numbers 1-15 one hundred percent of the time on 131,072-byte sub-blocks, and with the exclusion of test number 8 one hundred percent of the time on 16,384-byte packets. We have listed test number 8 as non-applicable, “NA”, on the 16,384-byte sub-blocks as test number 8 requires input files to be in excess of a 125,000-bytes.

DynAlysis’ conclusion is that the tested ciphers satisfy NIST statistical tests for randomness and produce the statistical performance of a commercial grade one-time pad cipher. Software speed optimizations and machine level coding will be required prior to microchip manufacturing.