<![CDATA[Polyglot persistence is the new approach for container-orchestrated microservices, where each microservice chooses a data-store technology that best fits its workload; and event-carried state transfer, where state changes are carried as immutable events through a shared log, instead of directly through synchronous calls between services. This literature survey investigates the interplay of three research lines: schema evolution in heterogeneous NoSQL and polyglot data stores, event-driven integration patterns to minimize synchronous coupling and distributed transaction management by the saga pattern. The review is based on the distributed-systems theory, empirical microservice-migration studies, and schema-evolution measurement studies, to synthesize a conceptual framework that enables safe, independently evolvable microservice data layers through schema-registry-governed compatibility checking, event-carried state transfer, and compensation-based transaction management. The synthesis emphasizes the classic dilemma of consistency vs availability, as summarized by the CAP theorem, as well as the operational rigidity of production database schemas, and the flexibility that NoSQL stores with schema-on-read are offered but not guaranteed. Four conceptual diagrams and three comparative tables are used to summarize the reviewed material as an integrated architectural viewpoint, whereas two formal notations are used to summarize the consistency and transaction-compensation reasons that underlie the framework. The review finds that schema governance, compatibility contracts and compensating transactions are complementary mechanisms that do not compete and finds open challenges around tooling maturity, consistency on cross-store data and schema-level lifecycle management at the container-orchestration level.]]>
Copyrights © 2024
