The global trend to extend connectivity to many devices for everyday use lead to the emergence of Edge computing. A paradigm where data and services are gradually pushed from the cloud which allows edge-level data storage and processing, increasing responsiveness through lower communication delay and lower transmitted traffic volumes. To allow this evolution, more and more devices (mobile devices as well) at the edge need to be mobilized and aggregated to compensate in the vast amount of resources previously offered by could-based services. These devices are closely related to human activity and as such mobility is expected. As different levels of mobility are observed, from mostly static sensors to highly dynamic connected vehicles, resource provisioning in dynamic environments on consumer devices emerged as a challenge, from the network to the application layer. To address this challenge, we aim to research on highly self-aware software services design model considering its own capabilities, limitations, anticipated needs, and associated contextual properties that we refer to as cognitive services. The cognitive services are autonomously deployed and orchestrated on the Edge-Fog-Cloud nodes and can migrate to other nodes due to, e.g., resource usage efficiency reasons, or for fulfilling dynamic accuracy and latency requirements.
To succeed, we intend to address the following research challenges:
• Energy-aware resource provisioning in an heterogeneous mobile context
• Dynamic and self-learned cognitive services migration between nodes
• Mobility-aware churn tolerant device-to-device communication
We propose activities involving exchange of staff and PhD students, strengthening both groups in Norway and France and providing easier access to the other groups methods. It will also encourage research collaboration and trigger initiatives for long lasting collaboration through EU-funded projects.