Interface Coordination & Control in a large-scale System-of-Systems : An industrial case study.

Abstract: Background: In developing and maintaining a large-scale system-of-systems (SoS), interface coordination & control (ICC) among interdependent subsystems is crucial for evolving individual subsystems and the overall SoS. However, the available literature offers little or no support to guide ICC in practice. Objectives: To describe how ICC can be done in the context of a large-scale SoS.We aimed to explore and analyze Ericsson’s (an example of a typical large-scaleSoS) current process and practices. To identify any limitations of the process, we investigated the challenges faced by practitioners regarding ICC, their reasons, and implications. Furthermore, we researched and proposed improvement possibilities to overcome the identified challenges. Method: We conducted a case study research and used observations, interviews, focus groups, and archival analysis for data collection. We used coding techniques from grounded theory and descriptive statistics to analyze qualitative and quantitative data, respectively. Moreover, the tabulation technique was employed to identify improvement opportunities. Results: We found that ICC is part of an overall technical coordination process. ICC is implemented through a recurrent cycle of activities. It brings together stakeholders to present upcoming changes, discuss their impact, and plan mitigation strategies. The outcome of ICC is a list of agreed-upon changes, impacted subsystems, and version numbers of interacting objects for the next release. We categorised the identified challenges in three groups: "people & subsystems", "tools, artifacts & technique", and "organization & management". These challenges impeded the effectiveness of the ICC process, quality of the SoS, and overall efficiency of the ICC and development process. Need for training, tool support, and stronger governance were identified as possible improvements to address multiple categories of identified challenges. Conclusion: The studied ICC process is well adapted for the large-scale SoS context. It is reasonably effective in minimizing the impact of evolving subsystems on the functioning and quality of SoS (as evident from the defect report analysis in this thesis). It also helps in cost-saving through early detection of compatibility-related issues. This thesis describes an ICC process used in practice, a novel contribution to the software configuration management (SCM) literature. In addition, companies in similar contexts can use the ICC description. The described ICC process is generalizableas as it is agnostic of the underlying development technologies used in a company.