Grant: Discovery Grants Program - Individual

With the increasing popularity of the Internet of Things (IoT), pervasive computing, healthcare services, sensor networks, and mobile devices, a lot of data is being generated at the edge of the network. Currently, the cloud is the most viable solution for storing, processing, and managing this data. However, clouds are usually located far away meaning that they are not efficient for time-sensitive services. Furthermore, moving the huge volume of data from edge to cloud leads to higher bandwidth consumption. Since cloud data centers might be located across countries, moving data will likely raise regulatory, security, and privacy issues. For this reason, the concept of fog computing (also known as edge computing) has been introduced. The fog can be seen as an extension of the cloud lying within the proximity of the underlying nodes and therefore providing quick response to applications that require it. Fog nodes can also preprocess and filter data to reduce the required bandwidth. Finally, trimmed data can be sent to the cloud for further analysis and enhanced service provisioning. To date, fog computing and IoT are considered as key enabling technologies for future 5G mobile networks. As a result, proper planning, design, and management of fog computing environments is crucial for network providers to minimize their costs (both capital and operational) and for users to enjoy a good quality of experience. The main goal of this research program is to develop models, algorithms and software tools to plan, design, and manage fog networks. More precisely, we first want to evaluate how the concept of fog computing used in conjunction with the cloud can improve performance compared to a standalone cloud. Then, we will look at how providers can deploy and manage their fog architecture while dealing with various considerations such as cost, delay, storage, compute, etc. Different models, algorithms, and software tools will be developed to assist network operators and service providers in the planning, deployment and management phases of their networks. The developed models and algorithms will be based on various optimization techniques as well as machine learning for efficient autonomous planning, management, and orchestration. The outcome of this research is expected to make significant contributions in developing new algorithms and tools to plan, design and manage fog networks. It will have a significant economic impact for all network providers looking to deploy and/or manage a fog network. In terms of technological impact, a proper deployment and management of fog networks will allow the creation of new real-time applications and services from different verticals. Finally, this research program is also an excellent opportunity for students to acquire expertise in the fields of IoT and fog computing which are gaining significant momentum in industry and academia.