Most of the research works conducted in the SEPIA group address the issue of resource management in datacenters. In the following, we distinguish strategies which specifically target energy optimization in the datacenter (consumption and thermal effects), from works which address the improvement of virtualized datacenter consolidation (which is more general, but may also address energy saving). In a third category, we present works that focussed on the improvement of operating system support (at the level of a single server) in such environments.
Context Data centers are computing infrastructures that host most of the services available on the internet. As data centers may include thousands of servers, both their energy consumption and their carbon footprint are significant. This has led many research projects to focus on optimizing these two objectives. This internship takes place in the context of centralized (Cloud) and decentralized (Edge, Fog) infrastructures.
Various ideas emerge from research / R&D to reduce the impact of computing infrastructures, and these ideas must be evaluated.
Context Data centers are computing infrastructures that host most of the services available on the internet. As data centers may include thousands of servers, both their energy consumption and their carbon footprint are significant. This has led many research projects to focus on optimizing these two objectives. This PhD takes place in the context of centralized (Cloud) and decentralized (Edge, Fog) infrastructures.
Various ideas emerge from research / R&D to reduce the impact of computing infrastructures, and these ideas must be evaluated.
Le présent sujet de stage s’inscrit dans le cadre du projet européen SLICES, dont l’objectif est la création d’une Infrastructure de Recherche (IR) pour le traitement numérique de la donnée, allant du capteur connecté (IoT) au traitement de données (Cloud), en passant par les protocoles réseau. Cette IR, en gestation, sera composée, entre autres, de nœuds comme ceux présents sur Toulouse, sur G5k [1] et LocURa4IoT [2].
L’objectif du stage est de proposer et expérimenter plusieurs scénarios illustrant cette IR.
Context High Performance Computing usage is growing from climate science studies to chemical research. The increased impact of these computation opens the field of research on how to manage and reduce their energy consumption. In the NumPEx project we aim at developing state-of-the-art skills and infrastructures in the field of exascale computing. One of the pillars of NumPEx focuses on making exascale computing sustainable.
To make informed cluster-level scheduling decisions and to provide feedback to users, information on the whole infrastructure is needed.
Context There is an increasing interest in a new distributed ML paradigm called Federated Learning (FL)[La17], in which nodes compute their local gradients and communicate them to a central server. This centralized server then orchestrates rounds of training over large data volumes created and stored locally at a large number of nodes. This training procedure repeats until some criterion are met. This enables the participating nodes (e.g., IoT devices, mobile phones, etc) to protect their data and solve the data security and privacy issues imposed by law.
Context High Performance Computing usage is growing from climate science studies to chemical research. The increased impact of these computation opens the field of research on how to manage and reduce their energy consumption. In the NumPEx project we aim at developing state-of-the-art skills and infrastructures in the field of exascale computing. One of the pillars of NumPEx focuses on making exascale computing sustainable.
To make informed cluster-level scheduling decisions and to provide feedback to users, information on the whole infrastructure is needed.
Context High Performance Computing usage is growing from climate science studies to chemical research. The increased impact of these computation opens the field of research on how to manage and reduce their energy consumption. In the NumPEx project we aim at developing state-of-the-art skills and infrastructures in the field of exascale computing. One of the pillars of NumPEx focuses on making exascale computing sustainable.
To make informed cluster-level scheduling decisions and to provide feedback to users, information on the whole infrastructure is needed.
Context High Performance Computing usage is growing from climate science studies to chemical research. The increased impact of these computation opens the field of research on how to manage and reduce their energy consumption. In the NumPEx project we aim at developing state-of-the-art skills and infrastructures in the field of exascale computing. One of the pillars of NumPEx focuses on making exascale computing sustainable.
To make informed cluster-level scheduling decisions and to provide feedback to users, information on the whole infrastructure is needed.
Nature du projet Développement logiciel
Description du travail demandé Obtenir des informations dans le noyau linux est compliqué et couteux. Ces informations sur la charge du système, le trafic réseau, la gestion de la mémoire, … sont souvent uniquement accessibles au travers de fichiers systèmes.
eBPF (Extended Berkeley Packet Filter) est une technologie kernel (lancée dans Linux 4.x) qui permet aux programmes d’être exécuté sans devoir modifier le code source du kernel ni ajouter de modules supplémentaires.
