Skip to content

ISCC-UNIT Instance-Code#

IEP: 0009
Title: ISCC-UNIT Instance-Code
Author: Titusz Pan tp@iscc.foundation
Comments: https://github.com/iscc/iscc-ieps/issues/14
Status: DRAFT
Type: Core
License: CC-BY-4.0
Created: 2022-09-28
Updated: 2024-01-05

Note

This document is a DRAFT contributed as input to ISO TC 46/SC 9/WG 18. The final version is developed at the International Organization for Standardization as ISO/DIS 24138

1. General#

  1. The Instance-Code shall be a checksum for any kind of data regardless of its media type.
  2. The Instance-Code shall match identical files and indicate data transmission errors or data manipulation.
  3. The Instance-Code shall be the prefix of a merkle root of a merkle tree as illustrated in Figure 11.
  4. The cryptographic hash algorithm used for Instance-Code calculation shall be updated if a security weakness is discovered.
  5. An update shall be indicated by a higher version number of the Instance-Code header.

Figure 11 - Example of a merkle tree

Figure 11 - Example of a merkle tree

NOTE 1

Depending on security and uniqueness requirements, applications can generate and store Instance-Codes up to a length of 256 bits in steps of 32 bits. Longer variants of an Instance-Code are extensions of the shorter versions (which function only as checksums) and remain compatible while improving the security of the hash.

NOTE 2

The use of tree-based hashing supports efficient containment-proofs, verified streaming and verification of data from trusted and untrusted sources. A data delivery application can provide cryptographic proofs for partial data and streams such that the receiving application can incrementally verify chunks of data against a given Instance-Code.

2. Format#

The Instance-Code shall have the data format illustrated in Figure 12:

Figure 12 Data format of the Instance-Code

Figure 12 - Data format of the Instance-Code

EXAMPLE 1: 64-bit Instance-Code in its canonical form:

ISCC:IAAZ3NGA3HTIYUQD

EXAMPLE 2: 256-bit Instance-Code in its canonical form:

ISCC:IADZ3NGA3HTIYUQD3SGC737FF6S5KRTRXY5DEU7ANCEMVTT4MDS2OQY

3. Inputs#

The input for calculating the Instance-Code shall be the bytes of a file, without reference to their meaning or structure.

4. Outputs#

Instance-Code processing shall generate the following output elements:

  1. iscc: the Instance-Code in its canonical form (required);
  2. datahash: a cryptographic hash of the input file encoded as multihash (optional);
  3. filesize: the size of the input file in bytes (optional).

5. Processing#

An ISCC processor shall calculate the Instance-Code as follows:

  1. Apply the tree-based cryptographic hash function to the file input using the appropriate number of bytes for the output digest.
  2. Use the hash digest of output as the ISCC-BODY of the Instance-Code.

For further technical details see source-code in modules code_instance.py of the reference implementation.

NOTE

The merkle tree chunk size is defined by the BLAKE3 1 algorithm to be 1024 bytes.

6. Conformance#

An implementation of the Instance-Code algorithm shall be regarded as conforming to the standard if it creates the same Instance-Code as the reference implementation for the same data input.

NOTE

The ISCC reference implementation uses the open source BLAKE3 1 library for merkle tree hashing and appropriate use of this software will generate the same codes as the reference implementation.

7. Bibliography#

  1. O’Connor, J., Aumasson, J.P., Neves, S., Wilcox-O’Hearn, Z., BLAKE3: one function, fast everywhere. Version 20211102173700, accessed July 2022. Available at https://github.com/BLAKE3-team/BLAKE3-specs/blob/master/blake3.pdf ↩↩