Advanced search
Start date
Betweenand


Adaptive offload in a heterogeneous fog environment: a case study for the health area

Full text
Author(s):
José Rodrigues Torres Neto
Total Authors: 1
Document type: Doctoral Thesis
Press: São Carlos.
Institution: Universidade de São Paulo (USP). Instituto de Ciências Matemáticas e de Computação (ICMC/SB)
Defense date:
Examining board members:
Júlio Cezar Estrella; Raphael Yokoingawa de Camargo; Edmundo Roberto Mauro Madeira; Cláudio Fabiano Motta Toledo
Advisor: Jo Ueyama
Abstract

The number of people with healthcare needs, such as elderly, disabled, and patients with reduced mobility, has been increased considerably in the world in recent years. This increase has created a trend in research involving healthcare for these patients through IoT-based applications, especially when they are liberated from the hospital and returned home. However, healthcare applications are considered critical and intolerant of delays, which require real-time responses. In this context, it is common for traditional cloud-based architectures to experience disconnection from the leading network or variations in bandwidth and latency, which generate losses to remote services, especially healthcare services. Consequently, the newly created fog computing enables data sharing and management at network edges, bringing intelligent decentralized processing closer to the data source and users, which ensures low latency. This continuous evolution implies that software systems must become increasingly versatile and flexible, self-adapting to the context of operating and intermittent environments. The main goal of this doctoral research is to mitigate the use of resources in real-time systems in IoT environments. In this context, a self-adaptive architecture and an autonomous middleware based on feedback cycles for dynamic data offloading in Fog-to-Cloud and Indie Fog multi-layer fog architectures are proposed. For this, a new policy for dynamic data offloading has been defined. The publish-subscribe paradigm was used for reliable communication and MAPE-K as a reference for the autonomous cycle. A quantitative evaluation of the middleware was performed through the simulation of an smarthome with two health applications: (i) heart monitoring and (ii) body temperature control. Also, a real case study on the sensing of the users physical and emotional state was performed with the applications: (i) facial recognition of the user, (ii) detection of falls, and (iii) recognition of emotion by voice. The results of the quantitative and qualitative evaluation have shown considerable improvement in performance and mitigation of resource use. The middleware is promising for the development of self-adaptive systems in the context of real-time and critical systems. (AU)

FAPESP's process: 16/25865-2 - Intelligent middleware for Internet of Things: an approach based on monitoring the emotional and physical state of patients in smart homes
Grantee:José Rodrigues Torres Neto
Support Opportunities: Scholarships in Brazil - Doctorate