Design and implementation of a distributed high-performance architecture for IoT data ingestion

This document provides an account of the project developed by undergraduate student Elena Ferro during her 302-hour internship at Danfoss S.r.l., a company operating within the field of the Internet of Things (IoT). The project goal was to rewrite a legacy microservice responsible for receiving messages from IoT devices via network sockets and performing some processing on the data. The decision to develop this project was driven by the need to improve its efficiency, scalability and maintainability: as a matter of fact, the original implementation was no longer meeting the performance requirements of the company and its non-optimal coding approach was making it complicated to implement additional features. Consequently, the new implementation would have to be thought with extensibility in mind while also ensuring high performances. The first phase consisted of an analysis of the existing code to understand the limitations and potential improvements. This was followed by the design stage, during which significant changes were introduced; for instance, the legacy application was split into two separate microservices and a cache to store frequently accessed data was adopted. Finally, a performance comparison was made between the original and the new implementation, which highlighted a significantly lower response time from the latter.