Impacts of Collisions within Hashing Algorithms and Safeguards Data

Quality assurance of data can be achieved using a one-way mathematical process that utilizes a hashing algorithm to provide data integrity. Ensuring data integrity is invaluable for the secure generation, transmission, archiving, and reviewing of safeguards information. Hash functions are also used to provide data authentication, either by the use of a symmetric key algorithm to form a keyed hash or with a public key algorithm to form a digital signature. In a typical safeguard application, data is collected by various sensors, which either retain the data or transmit the information to a central collection point. In either case, the sensor (or an additional component tied to the sensor) provides an authentication signature to accompany the sensor data. This signature can be verified at any time either automatically or manually to verify the data origin as well as the data integrity. The International Atomic Energy Agency (IAEA) specifies the use of a National Institute of Standards and Technology (NIST)-accredited hashing algorithm to produce the signature. This signature is ideally SHA-1, but non-NIST accredited algorithms can be used with specific agency approval. A hashing algorithm is useful only if it is highly unlikely for two messages to produce the same hash. Historically, some of the common algorithms in use have proved to be subject to collision; that is, two differing messages are found to produce the same message digest. For example a message M and a message M’ exist where the hash of (M) = H and the hash of (M’) also = H. At the CRYPTO2004 conference Wang et ali announced collisions for the MD4, MD5, HAVAL-128, and RACE Integrity Primitives Evaluation Message Digest (RIPEMD) hashing algorithms. More recently a theoretical collision was announced within the SHA algorithm. The impact of these collisions within the differing applications used for integrity and authentication is still not widely understood, nor are the long- and short-term ramifications on safeguards data. The impact of collisions to safeguards data as it is currently authenticated will be explored within this paper. This paper will also introduce future solutions to the issue of data authentication.