Abstract:
There are considerable amount of architectures and implementations for home automation systems in both industry and academia. Many of them target to inform home residents who can be anywhere in the world using Cloud technologies. In addition, some of them could be named as passive systems which only monitor events, environ mental conditions and inform the owner to take actions. This traditional approach is changing in the way of “smart things” with their increased processing capabilities. They are becoming active appliances which can make decisions by analysing the current state of the other appliances in the environment. In this study, we designed and implemented a well-structured home automation system, namely FOGHA, which enables both remote and local management of the home environment by separating tasks into layers and modules. Since direct communication between home appliances and Cloud servers is not desirable because of reliability, security and orchestration concerns, we divided the computation nodes into two main subsystems as Cloud and Fog. The latter subsystem is physically near to smart things and local orchestration of home environment is provided via Fog managers. In our system, Raspberry Pi (RPI) single board computers are used for this purpose in a distributed and fault-tolerant manner. Thus, they can be seen as the local brains of the overall system. While the gateway Fog manager additionally orchestrates other Fog managers and communicates with the Cloud, others collect data from smart things and manage them. Additional to the system design, we implemented a prototype of the FOGHA architecture and performed functional tests on it.