Serverless computing, mostly powered by Function-as-a-Service (FaaS), has been a growing field for several years now. In addition, the field of Edge computing has seen simultaneous growth as well, especially in sectors such as mobile computing, Internet of Things and Industry 4.0. Both approaches come with various advantages, although they are sometimes quite contrary to each other.
Now, a combination of the two paradigms for FaaS might be feasible to achieve to exploit their respective advantages. Yet, implementations for heterogeneous edge-cloud FaaS-environments have not yet been attempted, although the groundwork has been laid by edge-only FaaS-approaches such as tinyFaaS or Lambda@Edge.
In this thesis, an architecture for heterogeneous edge-cloud infrastructures based on technologies such as KubeEdge has been designed and implemented, as well as an automation to reproducibly deploy the infrastructure and all its components.
As a result, the infrastructure enables running a serverless computing framework such as OpenFaaS on a Kubernetes cluster consisting of both cloud and edge nodes whilst leveraging knowledge of, among other factors, physical node placement to optimize FaaS workload scheduling between cloud and edge nodes.
«
Serverless computing, mostly powered by Function-as-a-Service (FaaS), has been a growing field for several years now. In addition, the field of Edge computing has seen simultaneous growth as well, especially in sectors such as mobile computing, Internet of Things and Industry 4.0. Both approaches come with various advantages, although they are sometimes quite contrary to each other.
Now, a combination of the two paradigms for FaaS might be feasible to achieve to exploit their respective advanta...
»