Applications in the Cloud Edge Continuum: Environment, Challenges and Directions

Continuum Computing is a recent term born to refer to the capacity to smoothly integrate Cloud, Fog, and Edge Computing to let applications run wherever they can better exploit the characteristics of the infrastructure to perform better, faster, or more efficiently. Computing at the Continuum is not trivial for many reasons. First, because applications are not thought to be Continuum native, but just Cloud or Edge native, and second, because a Continuum infrastructure is composed of heterogeneous and highly distributed nodes, but applications need infrastructure transparency to work properly everywhere in the Continuum. In this regard, the use of containers and orchestrators helps, but those latter are usually heavy and unsuitable for high-constrained Edge devices. Moreover, internet use is huge in highly decoupled infrastructure and even transmitting sensitive data; therefore, keeping Continuum secure is challenging. All those conditions and challenges make the Continuum hard to define, standardize, and implement. In this thesis, we will go through all those aspects that make the realization of Continuum Computing hard, and for all of them, we will propose one or more suitable solutions that will be extensively discussed, designed, and tested. Ultimately, we will even analyze scenarios where Continuum Computing is needed and how our work can help satisfy those needs.