NextGenEMO: an Efficient Provisioning of Edge-Native Applications
Baudouin Herlicq, Abderaouf Khichane, Ilhem Fajjari · 2022
For years, Cloud infrastructures have driven the network services growth, providing increased scalability, enhanced performance and cost-effective resources. Nevertheless, the emergence of demanding 5G use cases poses several challenges that today’s Cloud infrastructures, despite their agility, struggle to overcome. In this context, Edge Computing has emerged to cope with Cloud’s shortcomings. This augmented architecture provides nearby resources, performing analytics tasks and data storage. However, it raises new challenges in terms of application design and management. In this perspective, the orchestration is the cornerstone that will enable the exploitation of edge infrastructures and the life-cycle management of multi-component applications. In this paper, we address the orchestration of edge-native applications issue in a practical manner. Specifically, we design and implement a multi-cluster orchestrator optimizing the provisioning of geo-distributed applications while considering the underlying edge infrastructure. To validate our approach, we evaluate the orchestration of our Towards5GS, an open-source edge-native 5G network service. Based on extensive experiments, the results obtained show that our solution achieved good results in terms of i) deployment time and ii) migration time.