A Pattern-Driven Middleware Architecture for IoT Data

Sensors in IoT systems are vital when it comes to collecting data across various applications automatically. However, there is no standardised software architecture for collecting and performing operations on data coming from IoT sensors, while also being able to aggregate other data sources. This is particularly evident in the context of managing asynchronous operations and streaming data, which are common in sensor-based systems. To address these issues, this paper proposes PD-MidI, a design pattern-based IoT middleware architecture that provides not only essential data collection, processing, and aggregation capabilities, but also incorporates comprehensive security, privacy, and anonymization features. We demonstrate how the middleware handles asynchronous data flows by implementing the Observer and Producer-Consumer design patterns, while also addressing access control requirements. To validate our proof-of-concept, we developed two use cases. The first use case demonstrated our middleware’s effectiveness in monitoring energy data from Hydro-Québec by collecting data from their energy meters and computing average energy consumption across various parameterisable time intervals. The second use case illustrated the middleware’s capability to collect and process real-time data from a Fitbit watch through a Firebase database. Our implementation results are promising and demonstrate the ability of our middleware to successfully provide flexibility, heterogeneity, and abstraction in data handling.