Improving Health Monitoring With Adaptive Data Movement in Fog Computing

Area of Science:

• Health Informatics
• Computer Science
• Data Management

Background:

• Pervasive sensing generates vast patient data for monitoring and analytics.
• Data quality and system performance issues can render this data unusable.
• Effective data management is crucial for reliable healthcare applications.

Purpose of the Study:

• To propose a data utility model for pervasive sensing in Fog environments.
• To address data quality and performance challenges impacting data usability.
• To enable utility-driven data management for improved analytics and patient monitoring.

Main Methods:

• Developed a context-dependent data utility model considering data quality (accuracy, completeness, consistency, timeliness) and Quality of Service (QoS) (availability, response time, latency).
• Proposed a goal-model-based approach to identify and address data quality and QoS violations.
• Evaluated the approach using a real-world dataset from a sensor-based physical activity monitoring project.

Main Results:

• The proposed model effectively assesses data utility based on application-specific requirements.
• The goal-model approach successfully identifies violations and suggests corrective actions.
• Demonstrated the practical applicability of the data utility model in a pervasive sensing scenario.

Conclusions:

• A novel data utility model enhances data management in Fog environments for pervasive sensing.
• Context-dependent quality assessment and QoS integration are key to ensuring data usability.
• The approach provides a framework for improving the reliability of data-driven healthcare applications.