Open Source Blockchain Software Health: A Theoretical Framework

Nijsse, Jeff
Litchfield, Alan
Johnson, Kenneth
Item type
Degree name
Doctor of Philosophy
Journal Title
Journal ISSN
Volume Title
Auckland University of Technology

The thesis combines research into blockchain software with the health of software ecosystems to produce a theoretical framework for open source blockchain health. The motivation comes from the blockchain trilemma, which argues that a blockchain network can achieve only two out of the three design objectives from decentralisation, security, and scalability. Between decentralisation and scalability, the objective is to investigate blockchain software from a health perspective to determine areas for improvement.

To fulfil these objectives, the study formulates a series of three focused research questions. The first is derived from the blockchain trilemma and asks: What are the factors that influence blockchain consensus? The second looks at software systems as a whole and asks: What is a definition of software health? The final research question unites software health and open source software (OSS) blockchain projects asking: What is included in a comprehensive model of OSS blockchain health?

To help answer these questions, hypotheses are formed that can then be tested within the overall methodology of Design Science Research (DSR). DSR allows for an iterative process where findings help inform processes and is particularly well suited for research in socio-technical fields. The focus of DSR is on the output artefacts that can be disseminated for further assessment and contribution to the field.

Among its contributions, the study offers a re-characterisation of the blockchain trilemma, distilling it into a dilemma that navigates between consensus methods, which are tied to decentralisation, and performance factors correlated with scaling capabilities. This, through the first question leads to the artefact of a taxonomy of blockchain consensus methods that reveals the landscape of algorithms underpinning decentralised networks.

In addition, the study introduces a conceptualisation of software health, drawn from literature in natural, business, and software ecosystems. It posits that the health of a software ecosystem is a composite construct comprising sustainability, robustness, and niche occupation. Sustainability is further decomposed into interest and engagement.

To operationalise the definition of health the study employs exploratory factor analysis to search for latent constructs from specific metrics identified in the literature. General Interest is gauged through the observed variables of forks, stars, and mentions, while Developer Engagement consists of commits, pull requests, comments, and contributors. Software Robustness is measured using the metrics criticality, time since the last update, geographic distribution of contributors, and market capitalisation ranking. These metrics are empirically substantiated through confirmatory factor analyses.

Structural equation modelling is used to add a structural element to the latent factors. The study follows a framework for developing theory in Information Systems to derive a theoretical framework for software health. The framework is the prime artefact of the work, not only advancing scholarly discourse and contributing to knowledge, but also yielding actionable guidance applicable to a variety of stakeholders, ranging from project managers to volunteer open source developers.

In summary, the study contributes to both academic and practical realms by providing a methodologically rigorous, empirically substantiated framework, and metrics for assessing the health of blockchain projects in the OSS ecosystem.

Publisher's version
Rights statement