CA2: Practical Archival and Compression of Covering Arrays

Abstract

Covering arrays (CAs) are combinatorial structures that play a significant role in software testing. While a sizable body of research has been dedicated to the generation of CAs, virtually no attention has been directed to their efficient storage, archival, and retrieval. Considering the potential economic advantages offered by the use of these structures, this is surprising: While the algorithmic complexity of most known methods used in their construction tends to be high, leading to expensive computations, the cost of storage has decreased in recent years. It thus seems prudent to identify and tackle the practical issues of large-scale and long-term archival and compression of such objects. This paper introduces CA2, the Compressed Covering Array Archive format, which offers competitive compression combined with CA-specific metadata that allows for targeted retrieval based on input specifications given in one of a variety of popular formats. An open source prototype implementation is provided as a utility for practitioners and building block for future research.