Federated cloud-native service placement in latency-sensitive and resource-constrained scenarios

This paper constitutes the contribution on the 5G-EPICENTRE service placement module. The module is a part of the federation layer and determines the best cluster for service deployment, where "best" is defined as the one contributing better to make an efficient (load balanced) use of computing resources, whilst fulfilling the stringent service requirements in terms of latency. In this direction, this paper presents a comprehensive solution to address the critical challenges associated with service management and resource allocation in emergency scenarios. This was achieved through the design and development of a built-in module, specifically for managing services based on their requirements in emergency situations. Our module ensures that services are deployed optimally, even under stringent emergency conditions. This paper introduces the context of the problem, formulates it, and evaluates it under a variety of conditions, including varying loads and cluster capacities, number of clusters, service requirements, or other parameters, more related with the algorithm design itself, such as number of services per service batch (i.e., number of services being deployed simultaneously in the emergency scenario).