Answers are grounded in this paper and cite their sources.
Choosing a software architecture is one of the most consequential decisions a software engineering team makes, yet in practice this choice is often driven by industry fashion rather than reasoned analysis.This paper examines three dominant architectural paradigms -the monolithic architecture, the modular monolithic architecture (colloquially the "minilith"), and the microservices architecture -and analyses their practical trade-offs across six engineering dimensions: deployment complexity, transactional consistency, independent scalability, fault isolation, cognitive load, and evolutionary migration potential.The central argument is that the modular monolithic architecture occupies an underused but practically valuable position in the design space, delivering the structural discipline of microservices without the distributed systems overhead those systems impose on teams that are not ready for it.A practitioner decision framework supported by three illustrative case scenarios is presented to guide architectural selection.The paper draws on peer-reviewed literature in software architecture, distributed systems, and organizational theory, and makes an explicit case against the trend of adopting microservices by default rather than by deliberate design.
No takes yet. Share an insight, caveat, or question.
Satish Chavali (2026) studied this question.