Aged information systems are commonly referred to as legacy information systems or just merely as legacy systems. Typically, these are mission-critical systems developed years ago that significantly resist evolution. Many organizations are confronted with these systems and consider them a burden because they cement their businesses and cause unreasonably high evolution and operating costs. However, they also cannot easily be discarded and replaced by modern solutions. Despite their relevance, no universally accepted concept exists that explains these systems or the mechanisms behind them. Accordingly, the properties and challenges associated with legacy systems vary significantly across different authors and their respective intentions – this needs to be revised, and this is where our research comes in. To this end, we first describe the causes of information systems’ aging, typical symptoms by which legacy systems are often recognized, and the consequences of using outdated solutions. Based on this empirical point of view, we introduce a holistic model to explain legacy systems objectively and introduce the concept of the business-technological age to distinguish between the chronological age and obsolescence of a deployed system. This approach provides practitioners and researchers the foundation for stringently explaining the hitherto fuzzy legacy phenomenon, promoting a common understanding of legacy systems, and simplifying communication through conceptualization. Practitioners can use this approach to better understand their stock application systems in terms of aging and improve their decision-making regarding evolution.

Big data has evolved to a ubiquitous part of today’s society. However, despite its popularity, the development and testing of the corresponding applications are still very challenging tasks that are being actively researched in pursuit of ways for improvement. One newly introduced proposition is the application of test driven development (TDD) in the big data domain. To facilitate this concept, existing literature reviews on TDD have been analyzed to extract insights from those sources of aggregated knowledge, which can be applied to this new setting. After introducing the different studies, lessons for the application of TDD in the big data domain are deducted and discussed. Finally, avenues for future works are proposed.

Big data is considered as one of the most promising technological advancements in the last decades. Today it is used for a multitude of data intensive projects in various domains and also serves as the technical foundation for other recent trends in the computer science domain. However, the complexity of its implementation and utilization renders its adoption a sophisticated endeavor. For this reason, it is not surprising that potential users are often overwhelmed and tend to rely on existing guidelines and best practices to successfully realize and monitor their projects. A valuable source of knowledge are use case descriptions, of which a multitude exists, each of them with a varying information density. In this design science research endeavor, 43 use cases are identified by conducting a thorough literature review in combination with the application and adaption of a corresponding template for big data projects. By a subsequent categorization, which is performed by identifying and employing a hierarchical clustering algorithm, nine different standard use cases emerge, as the contribution’s artifact. This provides decision-makers with an initial entry point, which can be utilized to shape their project ideas, not only by identifying the general meaningfulness of their potential big data project but also in terms of concrete implementation details.

Today, the amount and complexity of data that is globally produced increases continuously, surpassing the abilities of traditional approaches. Therefore, to capture and analyze those data, new concepts and techniques are utilized to engineer powerful big data systems. However, despite the existence of sophisticated approaches for the engineering of those systems, the testing is not sufficiently researched. Hence, in this contribution, a comparison of traditional software testing, as a common procedure, and the requirements of big data testing is drawn. The determined specificities in the big data domain are mapped to their implications on the implementation and the consequent challenges. Furthermore, those findings are transferred into six guidelines for the testing of big data systems. In the end, limitations and future prospects are highlighted.