Seamless Network Connectivity across Different Kubernetes Clusters

In the last two decades the cloud has gained a lot of importance, indeed the currenttrend is to engineer the new web applications to be cloud native, thus to be splitup in loosely-coupled micro-services, each one containerized and deployed as a partof a bigger application. The use of containers allows to cut oneself off the hostingphysical hardware and operating system, letting to focus on the main purposes of aweb application: to be widespread and high-available. The cloud allows to achievethis goal, by gathering the infrastructure control under the cloud provider tenantsand implementing the IaaS (Infrastructure as a Service) and PaaS (Platform asa Service) paradigms: the computational, networking and storage resources areprovided on demand to the cloud provider’s customers as if they were services. Atechnology that broke through the cloud market is Kubernetes, a project kicked offby Google in 2014 that allows to automate deployment, scaling, and managementof containerized applications. Beside the cloud, in recent years the edge computinghas gained a lot of importance: it is a distributed computing paradigm that bringsthe computational and storage resources close to the final user, in order to improvethe QoS standards in terms of latency and bandwidth.The goal of the project in which this thesis is involved is to create a federationof Kubernetes clusters that cooperate at the network edge: many different tenantsare connected together to cooperate in creating a federation of clusters withcomputational, storage and networking resources shared between them. In thisscenario every tenant can make its own resource cluster available to the federationby sharing or leasing them out in a federated environment.This solution needs a network model able to install a seamless network con-nectivity across different Kubernetesclusters. This work proposes a solution withminimal dependencies, which could be installed in a running cluster without bring-ing changes to configuration of the local environment. Permitting micro-services toconsume other services independently from the location where they are deployed