Design and development of the backend web architecture for knowledge marketplace mobile app
Zinkevich, Roman (2026)
2026
All rights reserved. This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-202603194600
https://urn.fi/URN:NBN:fi:amk-202603194600
Tiivistelmä
This thesis documents the design and implementation of the backend infrastructure for TutorSwap, a peer-to-peer knowledge marketplace designed to facilitate skill exchange among university students. The platform addresses the "Double Coincidence of Wants" in traditional barter systems by implementing a time-banking credit system, where one minute of teaching earns one credit that can be spent on learning other skills.
The technical objective was to establish a scalable, secure, and maintainable architectural foundation. The chosen technology stack includes Java with the Spring Boot framework, PostgreSQL for relational data integrity, and Docker for containerized deployment. A monolithic layered architecture was selected to balance development speed with system robustness, utilizing JWT (JSON Web Tokens) for stateless authentication and STOMP over WebSockets to enable real-time communication between users.
Key features implemented include a greedy matching engine for tutor discovery, a polymorphic messaging system to handle various data types, and an automated CI/CD pipeline via GitHub Actions. Evaluation of the system through unit and integration testing confirmed that the backend meets performance requirements, with API response times consistently under 400ms. The project concludes that while the current MVP provides a stable foundation, future iterations should explore microservices and dedicated search engines like Elasticsearch to support large-scale user growth.
The technical objective was to establish a scalable, secure, and maintainable architectural foundation. The chosen technology stack includes Java with the Spring Boot framework, PostgreSQL for relational data integrity, and Docker for containerized deployment. A monolithic layered architecture was selected to balance development speed with system robustness, utilizing JWT (JSON Web Tokens) for stateless authentication and STOMP over WebSockets to enable real-time communication between users.
Key features implemented include a greedy matching engine for tutor discovery, a polymorphic messaging system to handle various data types, and an automated CI/CD pipeline via GitHub Actions. Evaluation of the system through unit and integration testing confirmed that the backend meets performance requirements, with API response times consistently under 400ms. The project concludes that while the current MVP provides a stable foundation, future iterations should explore microservices and dedicated search engines like Elasticsearch to support large-scale user growth.