Uma plataforma para gerenciamento e aplicações em internet das coisas

• Autor(a) principal: Josué Batista Antunes
• Data de Publicação: 2016
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da UFMG
• Texto Completo: http://hdl.handle.net/1843/ESBF-ADZQV7
• topic: Internet das Coisas; Arquitetura de Gerenciamento; ManIoT; Internet das Coisas; Computação

The Internet of Things (IoT) paradigm encompasses a variety of devices, called things, with the ability to connect to the Internet and interact with each other to achieve common goals. Due to this ability to communicate, it is necessary to create solutions to integrate and manage the individual features of devices and provide new services, such as making use of data from a sensor to decide on an actuator. One of the challenges of IoT consists on the heterogeneity of the devices in the network, since each object may have different processing capabilities and/or communication standards. This requires a dynamic and context-aware configuration management system. However, the state of the art lacks management platforms that address the key issues of IoT in an integrated manner, such as multi-level management, context awareness and scalability. Furthermore, those works are restricted to a specific domain, and are therefore insufficient. This dissertation presents a platform that manages devices and applications on the Internet of Things, called ManIoT (Management for Internet of Things). It operates in local and global levels, and considers the perception of context. The division into two levels allows local control within a scenario (local level), while high-level policies are used to control multiple scenarios (global level). This dissertation also deals with the management of multiple applications and multiple users, treating problems such as conflicting commands and execution priority. A prototype of the platform was implemented, and experiments with different scenarios have shown that ManIoT supports applications and services that employ heterogeneous devices. The results indicate that ManIoT makes minimal use of computing resources and can thus be run in residential devices with modest processing capacity.