DRACeo : A smart simulator to deploy energy saving methods in a services/microservices based network
Résumé
Nowadays, there are many architectures for microservices deployment through the network (e.g. Docker-Kubernetes, Kalimucho, etc.) which aim at the creation of modular applications in the most efficient way. On the other hand, energy saving is an issue that concerns both sustainability and business global economy. For this reason, many researchers work to identify deployment solutions in order to save energy without decreasing functional QoS. In this work, we present DRACeo : a simulator to deploy and schedule microservices and their dependencies on various devices with software and hardware heterogeneity. DRACeo is able to apply various scheduling heuristics algorithms, including various philosophic management : centralized vs non-centralized in order to save energy without losing applications QoS. The simulator acts on moving, duplicating, starting/stopping microservices and devices. In addition, it performs important operations such as resource status monitoring, calculation of energy saved and current consumption (at device and network level). We have experimented some ideas based on our previous work (called Kaligreen) to demonstrate the effectiveness of DRACeo.
Origine : Fichiers produits par l'(les) auteur(s)