Legacy Enterprise Modernization Through Microservices Decomposition: A Seam-Based Extraction and Architecture Migration Framework

Abstract

Enterprises operating large, long-lived monolithic applications face a decomposition problem that is distinct from, and in several respects harder than, the problem of designing a new microservices system from scratch: a legacy monolith was not originally built with service boundaries in mind, its internal dependencies and shared database tables have accumulated over years of incremental feature work, and the organization cannot simply stop serving existing functionality while it is rebuilt. Abgaz et al.’s (2023) systematic review of monolith-to-microservices decomposition establishes that a substantial and still-growing body of techniques exists for identifying candidate service boundaries within an existing codebase, while Knoche and Hasselbring’s (2018) industrial account of using microservices for legacy software modernization documents that applying these techniques inside a live, continuously operating enterprise system introduces practical sequencing and risk-management challenges that boundary-identification techniques alone do not resolve. This paper proposes SEAM (Seam-Based Extraction and Architecture Migration), a framework in which a bounded-context boundary identification layer, an incremental strangler-style extraction layer, a data ownership separation layer, and a migration risk and rollback governance layer are coordinated so that a legacy monolith can be decomposed along its natural internal seams, one bounded context at a time, while continuing to serve production traffic throughout the migration. Following a design-science research approach, SEAM is, not empirically validated, through a comparative capability table, an  case scenario, and three  figures, with every quantitative value explicitly labeled as an  estimate rather than measured production data. The paper situates SEAM against the literature on monolith decomposition techniques, the evolutionary trajectory of microservices adoption, industrial legacy modernization practice, microservice design-pattern quality trade-offs, deployment cost comparisons, workload heterogeneity, and distributed consistency, and discusses the framework’s practical implications, assumptions, and limitations.