Research question and methodology
The central research questions are: "What are the key characteristics and objectives of a virtual monorepo approach?" and "To what extent can a virtual monorepo abstraction layer mimic the developer experience of a uniform monorepo workflow while operating on a distributed multi-repository codebase?"
These questions are particularly relevant given that the adoption of multi-repository architectures has tripled since 2018, while organizations struggle with fragmented workflows, high cognitive load, and repetitive operational tasks. The research employs a mixed-methods Design Science Research (DSR) approach, combining a multivocal literature review with quantitative performance experiments and qualitative case study observations.
Research design and techniques
The study follows a three-part approach based on design science research methodology. First, an extensive multivocal literature review was conducted, bringing together academic and industrial sources. This resulted in a conceptual framework that defines virtual monorepos based on five key characteristics: uniform git workflows, improved code visibility across repositories, simplified dependency management, coordinated change management, and insight into multi-repository build and CI/CD pipelines.
Based on this framework, MultiRepoHub was developed: a proof-of-concept browser extension that acts as an abstraction layer above GitLab repositories. The tool makes it possible to create virtual workspaces, group pull requests across multiple repositories, and perform cross-repository dependency analysis. The evaluation consisted of quantitative experiments measuring cloning time, commit latency, and pull request creation speed, supplemented by workflow step mapping and a week-long case study with experienced developers within the host organization.
Results: efficiency gains versus integration considerations
The results show a nuanced picture of the effectiveness of virtual monorepos. In terms of performance, MultiRepoHub demonstrated better scalability than manual multi-repository workflows. For example, cloning time increased to only 7.694 seconds for eight projects, compared to 19.726 seconds for sequential cloning. Commit operations showed even stronger improvements: the tool maintained stable performance of 0.457 seconds for eight repositories, compared to 1.099 seconds with traditional multi-repo scripts. This indicates superior scalability, despite an initial overhead. The most significant result was the reduction in procedural complexity: fragmented workflows were consolidated into virtually atomic commands, making the number of workflow steps virtually equal to that of native monorepos.
However, the qualitative case study revealed tensions between workflow simplification and engineering best practices. Developers appreciated the unified pull request overview and automated cross-repository change management functionality, but encountered issues with generic commit messages that sacrificed semantic depth for procedural efficiency. Although the tool reduced cognitive load and increased immediate task efficiency, its benefits proved to be highly dependent on deep integration with specific development environments and team agreements.
Implications and considerations
The research shows that a virtual monorepo is a viable middle ground for teams that want to leverage monorepo benefits without the costs and risks of a large-scale migration. The five-point framework developed in this research serves as a diagnostic tool for practitioners to evaluate existing multi-repository tooling and identify specific friction points. The findings show that the greatest return on investment is achieved by focusing on a uniform code review experience, as fragmented pull request reviews proved to be the biggest pain point for developers.
At the same time, the study points to a fundamental trade-off in abstraction level: workflow simplification is valuable, but must preserve semantic integrity and remain aligned with established engineering practices, such as meaningful commit conventions. The study concludes that the future of developer tooling does not lie in a binary choice between repository architectures, but in intelligent, context-aware abstraction layers that carefully balance automation with domain-specific requirements.
