arXiv subjects

Frédéric Le Mouël

Publications and source records attributed to Frédéric Le Mouël.

At least 19 recordsLinked to original sources

arXiv2026

Measuring Sustainability in Multi-Scale High-Performance Computing

The transition from traditional High Performance Computing (HPC) to the Computing Continuum emphasizes efficient resource management and sustainable practices across Multi-Scale hybrid architectures. This paper introduces a multidimensional metric framework to characterize these systems and guide deployment strategies for modern workloads. The framework combines Architectural Performance metrics (such as Throughput, Latency, Scalability), System Utilization, and key Sustainability and Accuracy indicators (such as Energy Efficiency and Power Consumption). Using a modular hybrid testbed, experiments reveal complex relationships among metrics, especially the trade-offs between accuracy and energy, and the efficiency of hybrid nodes. The guidelines help identify optimal operating points and lay the groundwork for improving orchestrators and schedulers (e.g., Kubernetes) to assign demanding applications, including AI and Quantum Computing, to suitable system modules, ensuring high performance and sustainability.

cs.DC↗

arXiv2026

Energy-Aware Quantum-Enhanced Computing Continuum

We discuss a Quantum-Enhanced Computing Continuum, a heterogeneous, hybrid architecture that integrates quantum processing units (QPUs) within an Edge-Cloud-HPC fabric. Promote sustainability by shifting from performance to "energy-aware integration.' The architecture has three layers: a Physical Layer with shared fiber-optic infrastructure, a Control and Orchestration Layer managed by the user, and an Application Layer with an Adaptive Quantum Classical Fusion (AQCF) framework. Tighter system integration, like moving from cloud coupling to cryogenic logic, reduces energy waste and "thermal footprints.' The aim is a Green Performance Advantage: energy per problem solved in the era of Advanced Computing.

cs.ET↗

arXiv2024

Usability and Performance Analysis of Embedded Development Environment for On-device Learning

This research empirically examines embedded development tools viable for on-device TinyML implementation. The research evaluates various development tools with various abstraction levels on resource-constrained IoT devices, from basic hardware manipulation to deployment of minimalistic ML training. The analysis encompasses memory usage, energy consumption, and performance metrics during model training and inference and usability of the different solutions. Arduino Framework offers ease of implementation but with increased energy consumption compared to the native option, while RIOT OS exhibits efficient energy consumption despite higher memory utilization with equivalent ease of use. The absence of certain critical functionalities like DVFS directly integrated into the OS highlights limitations for fine hardware control.

cs.SE↗

arXiv2023

Radio-Enabled Low Power IoT Devices for TinyML Applications

The proliferation of Internet of Things (IoT) devices has intensified the demand for energyefficient solutions supporting ondevice and distributed learning applications. This re search presents a circumscribed comparative analysis of radioenabled ultralow power IoT devices, specifically focusing on their suitability for computationheavy use cases. Our analysis centers on middleend IoT devices that serve as a vital interface between the Electronics and Machine Learn ing communities. The evaluation encompasses a diverse range of IoT hardware equipped with inte grated radios. We established functional datasheetbased criteria completed with accessibility and communitywise criteria to study and offer valuable insights into each selected node's performance tradeoffs, strengths, and weaknesses. This study provides crucial guidance for TinyML practitioners seeking to make informed device selections for their applications.

cs.NI↗

arXiv2021

Datacentric analysis to reduce pedestrians accidents: A case study in Colombia

Since 2012, in a case-study in Bucaramanga-Colombia, 179 pedestrians died in car accidents, and another 2873 pedestrians were injured. Each day, at least one passerby is involved in a tragedy. Knowing the causes to decrease accidents is crucial, and using system-dynamics to reproduce the collisions' events is critical to prevent further accidents. This work implements simulations to save lives by reducing the city's accidental rate and suggesting new safety policies to implement. Simulation's inputs are video recordings in some areas of the city. Deep Learning analysis of the images results in the segmentation of the different objects in the scene, and an interaction model identifies the primary reasons which prevail in the pedestrians or vehicles' behaviours. The first and most efficient safety policy to implement-validated by our simulations-would be to build speed bumps in specific places before the crossings reducing the accident rate by 80%.

cs.AI↗

arXiv2020

Infrastructure de Services Cloud FaaS sur noeuds IoT

In this article, we describe the PyCloudIoT cloud infrastructure. PyCloudIoT uses a FaaS cloud computing model for offloading numerical computations on a cluster with resource-constrained nodes rather than powerful datacenter. This infrastructure aims at exploiting unused resources of IoT nodes - already deployed at the edge of the network - to reduce latency of user requests. This extra computation must be done without significant energy consumption - IoT nodes being battery-powered. -- Dans cet article, nous décrivons l'infrastructure cloud PyCloudIoT. PyCloudIoT s'appuie sur un modèle de cloud computing FaaS pour du calcul numérique déporté vers une ferme de calcul composée de noeuds à capacités restreintes au lieu d'un datacentre puissant. Cette infrastructure vise à tirer profit des ressources inutilisées déployées sur nœuds IoT sans augmenter significativement leur consommation d'énergie et, en même temps, à rapprocher ces ressources des utilisateurs pour réduire la latence, en développant un modèle cloud en bord de réseau.

cs.DC↗

arXiv2017

Decentralized, Robust and Efficient Services for an Autonomous and Real-time Urban Crisis Management

The globalization of trade and the organization of work are currently causing a large migratory flow towards the cities. This growth of cities requires new urban planning where digital tools take a preponderant place to capture data and understand and decide in face of changes. These tools however hardly resist to natural disasters, terrorism, accidents, etc. Based on the expertise of the CITI laboratory of INSA Lyon and SC3 of the Industrial University of Santander, we propose to create the ALERT project - Autonomous Liable Emergency service in Real Time - with decentralized, reliable and efficient services, physically close to the citizens, taking decisions locally, in a relevant manner without risk of disconnection with a central authority. These information gathering and decision-making will involve the population with participatory and social approaches.

cs.CY↗

arXiv2016

Automated Application Offloading through Ant-inspired Decision-Making

-The explosive trend of smartphone usage as the most effective and convenient communication tools of human life in recent years make developers build ever more complex smartphone applications. Gaming, navigation, video editing, augmented reality, and speech recognition applications require considerable computational power and energy. Although smart- phones have a wide range of capabilities - GPS, WiFi, cameras - their inherent limitations - frequent disconnections, mobility - and significant constraints - size, lower weights, longer battery life - make difficult to exploiting their full potential to run complex applications. Several research works have proposed solutions in application offloading domain, but few ones concerning the highly changing properties of the environment. To address these issues, we realize an automated application offloading middleware, ACOMMA, with dynamic and re-adaptable decision-making engine. The decision engine of ACOMMA is based on an ant- inspired algorithm.

cs.DC↗

arXiv2016

Spontaneous Proximity Clouds: Making Mobile Devices to Collaborate for Resource and Data Sharing

The base motivation of Mobile Cloud Computing was empowering mobile devices by application offloading onto powerful cloud resources. However, this goal can't entirely be reached because of the high offloading cost imposed by the long physical distance between the mobile device and the cloud. To address this issue, we propose an application offloading onto a nearby mobile cloud composed of the mobile devices in the vicinity-a Spontaneous Proximity Cloud. We introduce our proposed dynamic, ant-inspired, bi-objective offloading middleware-ACOMMA, and explain its extension to perform a close mobile application offloading. With the learning-based offloading decision-making process of ACOMMA, combined to the collaborative resource sharing, the mobile devices can cooperate for decision cache sharing. We evaluate the performance of ACOMMA in collaborative mode with real benchmarks Face Recognition and Monte-Carlo algorithms-and achieve 50% execution time gain.

cs.NI↗

arXiv2015

Towards a Decoupled Context-Oriented Programming Language for the Internet of Things

Easily programming behaviors is one major issue of a large and reconfigurable deployment in the Internet of Things. Such kind of devices often requires to externalize part of their behavior such as the sensing, the data aggregation or the code offloading. Most existing context-oriented programming languages integrate in the same class or close layers the whole behavior. We propose to abstract and separate the context tracking from the decision process, and to use event-based handlers to interconnect them. We keep a very easy declarative and non-layered programming model. We illustrate by defining an extension to Golo-a JVM-based dynamic language.

cs.PL↗

arXiv2015

Real scenario and simulations on GLOSA traffic light system for reduced CO2 emissions, waiting time and travel time

Cooperative ITS is enabling vehicles to communicate with the infrastructure to provide improvements in traffic control. A promising approach consists in anticipating the road profile and the upcoming dynamic events like traffic lights. This topic has been addressed in the French public project Co-Drive through functions developed by Valeo named Green Light Optimal Speed Advisor (GLOSA). The system advises the optimal speed to pass the next traffic light without stopping. This paper presents results of its performance in different scenarios through simulations and real driving measurements. A scaling is done in an urban area, with different penetration rates in vehicle and infrastructure equipment for vehicular communication. Our simulation results indicate that GLOSA can reduce CO2 emissions, waiting time and travel time, both in experimental conditions and in real traffic conditions.

cs.MA↗

arXiv2015

Opportunities for a Truffle-based Golo Interpreter

Golo is a simple dynamically-typed language for the Java Virtual Machine. Initially implemented as a ahead-of-time compiler to JVM bytecode, it leverages invokedy-namic and JSR 292 method handles to implement a reasonably efficient runtime. Truffle is emerging as a framework for building interpreters for JVM languages with self-specializing AST nodes. Combined with the Graal compiler, Truffle offers a simple path towards writing efficient interpreters while keeping the engineering efforts balanced. The Golo project is interested in experimenting with a Truffle interpreter in the future, as it would provides interesting comparison elements between invokedynamic versus Truffle for building a language runtime.

cs.PL↗

arXiv2015

Semantic Service Substitution in Pervasive Environments

A computing infrastructure where everything is a service offers many new system and application possibilities. Among the main challenges, however, is the issue of service substitution for the application execution in such heterogeneous environments. An application would like to continue to execute even when a service disappears, or it would like to benefit from the environment by using better services with better QoS when possible. In this article, we define a generic service model and describe the equivalence relations between services considering the functionalities they propose and their non functional QoS properties. We define semantic equivalence relations between services and equivalence degree between non functional QoS properties. Using these relations we propose semantic substitution mechanisms upon the appearance and disappearance of services that fits the application needs. We developed a prototype as a proof of concept and evaluated its efficiency over a real use case.

cs.SE↗

arXiv2015

Router deployment of Streetside Parking Sensor Networks in Urban Areas

The deployment of urban infrastructure is very important for urban sensor applications. In this paper, we studied and introduced the deployment strategy of wireless on-street parking sensor networks. We defined a multiple-objective problem with four objectives, and solved them with real street parking map. The results show two sets of Pareto Front with the minimum energy consumption, sensing information delay and the amount of deployed routers and gateways. The result can be considered to provide urban service roadside unit or be taken into account while designing a deployment algorithm.

cs.NI↗

arXiv2014

How to Choose the Relevant MAC Protocol for Wireless Smart Parking Urban Networks?

Parking sensor network is rapidly deploying around the world and is regarded as one of the first implemented urban services in smart cities. To provide the best network performance, the MAC protocol shall be adaptive enough in order to satisfy the traffic intensity and variation of parking sensors. In this paper, we study the heavy-tailed parking and vacant time models from SmartSantander, and then we apply the traffic model in the simulation with four different kinds of MAC protocols, that is, contention-based, schedule-based and two hybrid versions of them. The result shows that the packet interarrival time is no longer heavy-tailed while collecting a group of parking sensors, and then choosing an appropriate MAC protocol highly depends on the network configuration. Also, the information delay is bounded by traffic and MAC parameters which are important criteria while the timely message is required.

cs.NI↗

arXiv2014

VANET Applications: Hot Use Cases

Current challenges of car manufacturers are to make roads safe, to achieve free flowing traffic with few congestions, and to reduce pollution by an effective fuel use. To reach these goals, many improvements are performed in-car, but more and more approaches rely on connected cars with communication capabilities between cars, with an infrastructure, or with IoT devices. Monitoring and coordinating vehicles allow then to compute intelligent ways of transportation. Connected cars have introduced a new way of thinking cars - not only as a mean for a driver to go from A to B, but as smart cars - a user extension like the smartphone today. In this report, we introduce concepts and specific vocabulary in order to classify current innovations or ideas on the emerging topic of smart car. We present a graphical categorization showing this evolution in function of the societal evolution. Different perspectives are adopted: a vehicle-centric view, a vehicle-network view, and a user-centric view; described by simple and complex use-cases and illustrated by a list of emerging and current projects from the academic and industrial worlds. We identified an empty space in innovation between the user and his car: paradoxically even if they are both in interaction, they are separated through different application uses. Future challenge is to interlace social concerns of the user within an intelligent and efficient driving.

cs.CY↗

arXiv2014

Performance Comparison of Contention- and Schedule-based MAC Protocols in Urban Parking Sensor Networks

Network traffic model is a critical problem for urban applications, mainly because of its diversity and node density. As wireless sensor network is highly concerned with the development of smart cities, careful consideration to traffic model helps choose appropriate protocols and adapt network parameters to reach best performances on energy-latency tradeoffs. In this paper, we compare the performance of two off-the-shelf medium access control protocols on two different kinds of traffic models, and then evaluate their application-end information delay and energy consumption while varying traffic parameters and network density. From the simulation results, we highlight some limits induced by network density and occurrence frequency of event-driven applications. When it comes to realtime urban services, a protocol selection shall be taken into account - even dynamically - with a special attention to energy-delay tradeoff. To this end, we provide several insights on parking sensor networks.

cs.NI↗

arXiv2013

Modèle multi-échelles pour les services dans les VANET

Service deployment, integration and interactions in VANET are fast expanding. They are now focusing only on driver security. In the future, with in-car application deployments, the car will be considered as a user extension - in the same way as smartphones actually. We propose a multi-level graph model to challenge this VANET / Application / User integration.

cs.NI↗