Fog Computing in Smart Home Environments based on Microsoft Azure IoT Technology

Abstract

In this thesis, an effort has been made to develop an experimental fog computing system then afterwards to test multiple issues such as privacy and security concerns. The tested issues indeed include the device management patters in this fast-pacing developing area of study. Therefore, getting started issues stand in front of developers are important problems to solve. Some of the research issues that emerged early in the development process are worth mentioning such as development machine used compatibility, edge device choosing decisions, and relying on current network infrastructures regarding offloading pre-developed workloads, and sending/receiving of telemetry. Many of open-source components, technologies, and tools have been used in the development of the experimental fog computing system. A smart home scenario has been used as an analogy to develop and run the system along with simple device management principles and patterns leveraged to make the execution a reality as possible. Primarily, and unlike traditional ways, the development machine should be prepared for the development process with original OS license along containers engine software. The edge device appropriately chosen to run Raspbian OS with container engine software as with development machine. .NET Core should be appropriately configured in both the development machine and the edge device. However, development and deployment of modules (also known as image, containers) by VS Code and Docker container engine and pushing the modules to cloud storage is the main and first step of the total architectural design. Afterwards, and after the edge device get notified of the manifest of the first step, the pulling of images is starting by the client-side application, the runtime of the edge device, to be used and that is the second step of the architectural design. Results from the operation of the developed experimental system formed a clear evidence to the assumptions made in this thesis when it had been still in the development process. The main and basic aspects of rebooting, factory resetting, v and configuring management patterns have been applied to take related action whenever some related circumstances met. Also, the overlapping of work between these patterns is seen to be mandatory to consider when developing own patterns. Finally, and to make the developed platform seems to be an end-to-end system, various operational cases have been studied to experimentally verify the difference when enabling edge analytics and when send raw readings as generated to the cloud. Indeed, the case of symmetric encryption of sent messages to increase the security aspects of the application has been studied and experimentally verified.