Distributed Computing Framework Based on Software Containers for Heterogeneous Embedded Devices

Thumbnail Image
Journal Title
Journal ISSN
Volume Title
Perustieteiden korkeakoulu | Master's thesis
Service Design and Engineering
Degree programme
Master's Programme in ICT Innovation
58 + 0
The Internet of Things (IoT) is represented by millions of everyday objects enhanced with sensing and actuation capabilities that are connected to the Internet. Traditional approaches for IoT applications involve sending data to cloud servers for processing and storage, and then relaying commands back to devices. However, this approach is no longer feasible due to the rapid growth of IoT in the network: the vast amount of devices causes congestion; latency and security requirements demand that data is processed close to the devices that produce and consume it; and the processing and storage resources of devices remain underutilized. Fog Computing has emerged as a new paradigm where multiple end-devices form a shared pool of resources where distributed applications are deployed, taking advantage of local capabilities. These devices are highly heterogeneous, with varying hardware and software platforms. They are also resource-constrained, with limited availability of processing and storage resources. Realizing the Fog requires a software framework that simplifies the deployment of distributed applications, while at the same time overcoming these constraints. In Cloud-based deployments, software containers provide a lightweight solution to simplify the deployment of distributed applications. However, Cloud hardware is mostly homogeneous and abundant in resources. This work establishes the feasibility of using Docker Swarm -- an existing container-based software framework -- for the deployment of distributed applications on IoT devices. This is realized with the use of custom tools to enable minimal-size applications compatible with heterogeneous devices; automatic configuration and formation of device Fog; remote management and provisioning of devices. The proposed framework has significant advantages over the state of the art, namely, it supports Fog-based distributed applications, it overcomes device heterogeneity and it simplifies device initialization.
Di Francesco, Mario
Thesis advisor
Di Francesco, Mario
internet of things, fog computing, software containers, docker swarm, embedded devices, lightweight virtualization
Other note