Storage follows the workflow
SQL Server brings together users, orders and billing; MongoDB holds product documents. Redis manages the cart and undo/redo, while Cassandra records session events.
CASE STUDYINTEGRATION & PERSISTENCE
Polyglot Persistence
Four engines. Distinct responsibilities.
Project · Local prototype · console
JavaSQLSQL ServerMongoDBRedisCassandraDockerDocker Compose
Overview
Four persistence engines, distinct responsibilities and one Java console workflow.
SQL Server, MongoDB, Redis and Cassandra support catalog, cart, order and billing workflows. A local integration prototype with no web UI or REST API.
Problem
Assign each operation to an appropriate storage model and bring the workflows together in one application.
My role
I worked on the application and its persistence integration, contributing to Cassandra session records and local-environment configuration.
Architecture
Console menu → business services → engine-specific DAOs. SQL Server stores users, orders and billing; MongoDB the catalog; Redis the cart and undo/redo; Cassandra session records. Docker Compose prepares the local environment.
Technical decisions
Engine-specific access
Services connect through DAOs specific to each engine. Integration runs in the local Docker Compose environment, without production-grade distributed transaction coordination.
Challenges and solutions
Integration requires connections and initialization for four engines, and consistent login/logout records. The application does not implement production-grade distributed transaction coordination.
Outcomes
Console workflows using four engines with distinct roles, with fixes to session records and local-environment configuration.