TL;DR
This paper introduces PRESS, a pervasive environment sensing and sharing framework designed to provide citizens with real-time, granular air quality information to encourage sustainable behavior. It leverages IoT, MQTT, and WebSockets, featuring novel congestion detection and recovery mechanisms for robust data delivery.
What Problem Does This Paper Address?
The paper tackles the challenge of increasing environmental and economic sustainability by making citizens more aware of their immediate surroundings, specifically air quality. The core problem is the lack of accessible, up-to-date, and contextually relevant environmental data. Existing solutions often struggle with integrating diverse data sources, managing large volumes of information, and ensuring reliable delivery, especially under varying network conditions. The authors argue that providing users with dynamic, granular information (e.g., home, district, city levels) about air pollution can empower them to make better decisions, influence personal habits, and advocate for more sustainable policies.
Methods
The PRESS framework is built upon several key technical components and design principles:
- Internet of Things (IoT) Paradigm: PRESS leverages IoT to collect data from a variety of heterogeneous sensors.
- PeRvasive Environment Sensing and Sharing (PRESS) Project: This project aims to increase user awareness of their environment through real-time, user-centric information.
- Heterogeneous Data Management: The framework is designed to handle diverse data types, from quantitative sensor readings (air pollution) to qualitative user perceptions.
- Pluggable and Extensible Architecture: PRESS employs a flexible architecture that allows for the seamless integration of new data types and devices without requiring system recompilation or service interruption. This is achieved through a modular design and adherence to communication standards.
- Messaging Protocols:
- MQTT (Message Queue Telemetry Transport): Used for efficient, publish/subscribe-based communication between gateways and the server, particularly suited for resource-constrained IoT devices due to its low overhead and support for Quality of Service (QoS) levels.
- WebSockets: Employed for real-time, bidirectional communication between the server and the user interface, enabling push notifications of new data.
- Kura IoT Gateway: The framework utilizes the open-source Kura project for its field-side gateway module, providing tools for component management and MQTT communication.
- Spring Framework: The server-side components are built using the Spring framework, known for its flexibility and extensibility, particularly with the Spring Integration project for message handling.
- Congestion Detection and Recovery Mechanisms: A novel aspect of PRESS is its proactive approach to managing network traffic. This includes:
- Congestion Detection: Inspired by DFlow, it monitors end-to-end message delays to identify when message rates exceed processing capabilities. It uses conditions based on Lowest Delay of Long periods (LDL) and Lowest Delay of Short periods (LDS) to trigger alerts.
- Congestion Recovery: Implements a probabilistic message filtering solution that reduces the rate of non-priority messages during congestion, ensuring the delivery of critical data. This mechanism dynamically adjusts the probability of dropping regular packets based on the current network load.
Datasets
Not discussed in the paper. The paper focuses on the framework's design and evaluation using simulated and real-world sensor data, but does not detail specific datasets used for training or benchmarking beyond the types of environmental data collected.
Key Insights and Ideas
- User Awareness as a Sustainability Driver: The fundamental insight is that empowering citizens with accessible, real-time environmental data is a powerful lever for driving behavioral change towards sustainability.
- Synergy of MQTT and WebSockets: The paper highlights the effectiveness of combining MQTT for efficient device-to-cloud communication and WebSockets for real-time cloud-to-user updates, creating a robust end-to-end data pipeline.
- Proactive Congestion Management is Crucial: For IoT systems dealing with potentially high volumes of heterogeneous data, simply relying on scalable protocols isn't enough. Novel mechanisms to detect and actively manage congestion are essential for maintaining service continuity and data quality.
- Granularity Matters: The ability to provide environmental data at different levels of detail (home, district, city) is key to making the information relevant and actionable for users.
- Extensibility is Paramount: In dynamic IoT environments, a framework must be easily extendable to incorporate new sensors and data types without significant re-engineering.
Key Results
- Scalability: The PRESS framework demonstrated good scalability, capable of managing a significant volume of data. The paper suggests it can handle over 400,000 devices with appropriate replication of the web-app module.
- Protocol Performance:
- WebSocket vs. REST: WebSocket generally offered lower latency and better performance for real-time data delivery compared to REST, which relies on polling.
- MQTT QoS Levels: While higher MQTT QoS levels (at-least-once, exactly-once) ensure better message delivery reliability, they introduce significant overhead, potentially reducing overall scalability compared to best-effort (QoS 0). Best-effort MQTT was found to be adequate for many scenarios, offering better scalability.
- Congestion Management Effectiveness: The implemented congestion detection and recovery mechanisms were shown to be effective. When enabled, they maintained end-to-end latency below 0.1s for most of the time, even under high load, preventing the significant delays and message queuing observed when these mechanisms were disabled. This ensures prompt delivery of priority messages and service continuity.
- Framework Flexibility: The use of a JSON configuration file for defining devices and topics allows for dynamic integration of new sensors without modifying the core framework, demonstrating its extensibility.
Limitations and Future Work
- Congestion Recovery Granularity: The paper notes that the congestion recovery mechanism's granularity can be tuned by adjusting the
Nmaxparameter. WhileNmax=10was used, further exploration of optimal values for different scenarios could be beneficial. - Persistence Bottleneck: The authors identify that the primary bottleneck in their current PRESS framework implementation is persistence management and database access.
- Future Work: The authors suggest integrating PRESS with non-relational NoSQL persistence storage solutions (like MongoDB) to address the identified bottleneck and further enhance scalability. They also mention ongoing investigations into how large packet payloads might impact performance in different application domains.
Summary updated