Scalability, heterogeneity, and device life-cycle are among the key concerns regardingInternet of Things (IoT) system development and deployment. Consequently,they require robust solutions for managing gateway-resident services asthe systems evolve while being deployed in a large scale.This thesis studies kernel-level virtualization in the context of designing andimplementing a software management system for use in IoT edge gateways. Amanagement system is implemented for a gateway which is deployed to severalhouseholds, where the gateways act as base stations for smart electricity plugs.The gateway is designed to be used in a private household network, connecting tothe Internet using the household’s own connection without modifications to routersettings and with no configuration required from the homeowner.The implementation runs on the GNU/Linux operating system on the RaspberryPi 2 single-board computer. It uses Docker as the virtualization system,Ansible for host-level configuration management and MQTT as the primary protocolfor communication with cloud-resident management services.The gateway was deployed to a number of households in the pre-pilot phase ofthe Flex4Grid project. During the pre-pilot, the implemented software managementsystem provided a convenient way of deploying and maintaining middlewarepackages running inside the gateways, as well as being capable of updating itselfas required.
|Publication status||Published - Aug 2016|
|MoE publication type||G2 Master's thesis, polytechnic Master's thesis|