Combinatorial Coverage Measurement of Test Vectors used in Cryptographic Algorithm Validation (Presentation)

Published

September 28, 2017

Author(s)

Dimitris Simos, Stavros Mekesis, D. Richard Kuhn, Raghu N. Kacker

Abstract

We measured the combinatorial coverage of test vectors provided by the NIST Cryptographic Algorithm Validation Program (CAVP). Input models were defined and test vectors measured and analyzed for 2-way, 3-way, and 4-way combinatorial coverage. The results of our measurement show that some test vectors do not achieve a full 2-way to 4-way combinatorial coverage, so we generated the missing combinations for these vectors and extended the test suites to achieve a full 2-way coverage. We also conducted differential testing on popular AES implementations, such as OpenSSL, LibreSSL, Crypto++, PyCrypto, using the extended test suites. Our differential testing of AES implementations on these test suites showed no discrepancies between the implementations. Finally, we use the NIST Cryptographic Algorithm Validation System (CAVS) as a golden system against which the AES implementations are tested.

Conference Dates

September 25-28, 2017

Conference Location

Gaithersburg, MD, US

Conference Title

IEEE Software Technology Conference

Pub Type

Conferences

Download Paper

Local Download

Keywords

AES, combinatorial testing, combinatorics, computer security, cryptography

Citation

Simos, D. , Mekesis, S. , Kuhn, D. and Kacker, R. (2017), Combinatorial Coverage Measurement of Test Vectors used in Cryptographic Algorithm Validation (Presentation), IEEE Software Technology Conference, Gaithersburg, MD, US, [online], https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=923635 (Accessed May 2, 2025)