Honey bees have been domesticated by humans for several thousand years and mainly provide honey and pollination, which is fundamental for plant reproduction. Nowadays, the work of beekeepers is constrained by external factors that stress their production (parasites and pesticides, among others). Taking care of large numbers of beehives is time-consuming, so integrating sensors to track their status can drastically simplify the work of beekeepers.

Precision Beekeeping is a subfield of Precision Agriculture. It collects data on colonies of bees using sensors to assist beekeepers and preserve bees. The existing solutions include collecting voluminous data, such as images and sound, from up close or inside a beehive. The collection, transfer, and processing of such data are energy-intensive processes. However, most Precision Beekeeping systems work under a limited energy budget. Thus, such systems must take this constraint into account. This research focuses on the placement of energy-aware Precision Beekeeping services at the edge and in the cloud. We deploy a set of smart beehives in order to conduct experiments and collect real data. Then, we simulate large-scale systems with embedded services to add more realism to the experimental study. Our results show that parameters like the number of smart beehives, the frequency of data collection, and the characteristics of the cloud server impact the placement of these services to enhance energy efficiency. Some experimental parts of this work occur in the CPER LECO/GreenCube project and some parts are financially supported by aivancity School for Technology, Business & Society Paris-Cachan. This work, within the the Ph.D. of Hugo Hadjur is co-advised by Doreid Ammar (Academic Dean and Professor at aivancity School for Technology, Business & Society Paris-Cachan and external member of Avalon team) and Laurent Lefevre 8.

Creating energy aware application with limited environnemtal impacts needs a complete redesign. Laurent Lefevre with some colleagues from the GDS EcoInfo group have explored the various facets of ecodesign. This has resulted to a new version of a brochure available for software developers. This brochure 18 has been downloaded several thousands of times since the publication of the first version.

With colleagues from the GDS EcoInfo group (Anne-Laure Ligozat, Dernis Trystram) we explore criteria for assessing the environmental impacts of responses to calls for projects involving Artificial Intelligence (AI) methods. When proposing these criteria, we take into account, in addition to the general impacts of digital services, the specificities of the AI field and in particular of machine learning: impacts of the learning and inference phases and data collection 17.

Efficient resource management is a challenge for Cloud service providers. They need to respond to growing demand by limiting the oversizing of their infrastructures, which can generate avoidable costs and environmental impact. Mechanisms derived from virtualization are used to optimize infrastructure sizing, such as elasticity and consolidation, but economic or physical constraints can hinder their adoption. In this work, we propose a consolidation method based on the observation of the resource consumption of virtual machines (VMs) in the infrastructure of a private Cloud host. With a suitable evaluation algorithm and using memory over-allocation mechanisms, we are able to generate sets of VMs positioned in such a way as to limit the number of physical machines used, while ensuring the quality of service required by users 14.

EcoIndex has been proposed to evaluate the absolute environmental performance of a given URL using a score ranging from 0 to 100 (the higher, the better). In this article, we make a critical analysis of the initial approach and propose alternatives that no longer calculate a plain score but allow the query to be situated among other queries. The generalized critiques come with statistics and rely on extensive experiments (first contribution). Then, we move on to low-cost Machine Learning (ML) approaches (second contribution) and a transition before obtaining our final results (third contribution). Our research aims to extend the initial idea of analytical computation, i.e., a relation between three variables, in the direction of algorithmic ML computations. The fourth contribution corresponds to a discussion on our implementation, available on a GitHub repository. Beyond computational questions, it is important for the scientific community to focus on this question in particular. We currently promote using well established ML techniques because of their potential. However, we also question techniques for their frugality or otherwise. Our data science project is still at the data exploration stage. We also want to encourage synergy between technical expertise and business knowledge because this is fundamental for advancing the data project 5.

Generative AI (Gen-AI) represents a major growth potential for the digital industry, a new stage in digital transformation through its many applications. Unfortunately, by accelerating the growth of digital technology, Gen-AI is contributing to the significant and multiple environmental damage caused by its sector. The question of the sustainability of IT must include this new technology and its applications, by measuring its environmental impact. To best respond to this challenge, we propose various ways of improving the measurement of Gen-AI's environmental impact. Whether using life-cycle analysis methods or direct measurement experiments, we illustrate our methods by studying Stable Diffusion a Gen-AI image generation available as a service. By calculating the full environmental costs of this Gen-AI service from end to end, we broaden our view of the impact of these technologies. We show that Gen-AI, as a service, generates an impact through the use of numerous user terminals and networks. We also show that decarbonizing the sources of electricity for these services will not be enough to solve the problem of their sustainability, due to their consumption of energy and rare metals. This consumption will inevitably raise the question of feasibility in a world of finite resources. We therefore propose our methodology as a means of measuring the impact of Gen-AI in advance. Such estimates will provide valuable data for discussing the sustainability or otherwise of Gen-AI solutions in a more transparent and comprehensive way 3.

The global energy demand for digital activities is constantly growing. Computing nodes and cloud services are at the heart of these activities. Understanding their energy consumption is an important step towards reducing it. On one hand, physical power meters are very accurate in measuring energy but they are expensive, difficult to deploy on a large scale, and are not able to provide measurements at the service level. On the other hand, power models and vendor-specific internal interfaces are already available or can be implemented on existing systems. Plenty of tools, called software-based power meters, have been developed around the concepts of power models and internal interfaces, in order to report the power consumption at levels ranging from the whole computing node to applications and services. However, we have found that it can be difficult to choose the right tool for a specific need. In this work, we qualitatively and experimentally compare several software-based power meters able to deal with CPU or GPU-based infrastructures. For this purpose, we evaluate them against high-precision physical power meters while executing various intensive workloads. We extend this empirical study to highlight the strengths and limitations of each software-based power meter. This research is a joint work with Denis Trystram (LIG Laboratory) and Anne-Cécile Orgerie (IRISA, Rennes) 9.

Data centers are very energy-intensive facilities that can generate various environmental impacts. Numerous energy, power, and environmental leverages exist and can help cloud providers and data center managers to reduce some of these impacts. But dealing with such heterogeneous leverages can be a challenging task that requires some support from a dedicated framework. This research presents a new approach for modeling, evaluating, and orchestrating a large set of technological and logistical leverages. Our framework is based on leverages modeling and Gantt chart leverages mapping. Some experimental results based on selected scenarios show the pertinence of the proposed approach in terms of management facilities and potential impacts reduction 2. The research is done inside the challenge (Défi) between Inria and OVHcloud company.

Today, a massive shift is ongoing in telecommunication networks with the emergence of softwarization and cloudification. Among the technologies which are assisting these shifts, one of them is NFV (Network Function Virtualization). NFV is the network architecture that decouples network functions from hardware devices (middleboxes) with the help of a virtual component known as VNF (Virtual Network Function). VNF has shifted the network technological paradigm. Before: Network Function was performed by physical equipment, and service providers acquired its property for the lifetime of the relying hardware (instead counted in years). Today, Network functions are software that service providers develop or acquire purchasing licenses. A license defines software's Right to Use (RTU).

Therefore, if licensing in NFV is not appropriately managed, service providers might (1) be exposed to counterfeiting and risk heavy financial penalties due to non-compliance; (2) might overbuy licenses to cover poorly estimated usages. Thus, mastering network function license through implementing Software Asset Management and FinOps (Finance and DevOps) is essential to control costs. In this research, our primary problem is to minimize the TCO (Total Cost of Ownership) of software cost (VNF), providing Quality of Services (QoS) to a specific amount of users. Software costs include various costs, from development to maintenance, integration to release management, and professional services. Our research focuses on proprietary software (developed by a publisher and sold via a paid license).

We considered that TCO consists of the software license cost, the resources necessary to execute and operate SW, and the energy consumed by this execution. In this research, first, we have identified the need for a standardized VNF licensing model, which is highly dependent on the VNF provider's creativity; This lack of standards places CSPs (Communication Service Providers) at risk of having to delegate the management of rights to suppliers. Hence, we proposed a licensing model based on the metrics, which help to quantify the usage of the VNF. After estimating the license of VNF, we estimated the license cost. Afterward, we presented several ways to minimize the license cost depending upon the different use cases, which depend on the user's scenario and needs. Then after, with the help of industrial knowledge, we found that reducing resource consumption to minimize the TCO providing QoS affects the deployment of the VNF directly or indirectly, which impacts the licensing. Thus, the licenses and resources are interdependent. We used these costs to construct the software's total cost. After that, we proposed several ways to reduce the software's total cost by fulfilling the client's requirements. Then after, we considered the energy and its associated cost of VNF. The energy consumption of the VNF is dependent on resource consumption, and resources usages impact the license. Thus, we can see that these three costs are interdependent: license, resources, and energy cost of VNF. Hence, we consider these costs and constructed TCO. Minimizing TCO fulfilling the client's requirements is challenging since it is a multi-parameter. Therefore, we proposed several heuristical algorithms based on resource sharing and consolidation to reduce the TCO depending on the license, resource preference, and the client's scenarios.

This work 16 was obtained during the PhD of Ghoshana Bista co-advised by Anne-Lucie Vion (Orange) and Eddy Caron.

With the rise of Data as a Service, companies understood that whoever controls the data has the power. The past few years have exposed some of the weakenesses of traditional data management systems. For example, application owner can collect and use data to their own advantage without the user's consent. We defined the SkyData concept, which revolves around autonomous data evolving in a distributed system. This new paradigm is a complete break from traditional data management systems. This paradigm is born from the many issues associated with traditional data management systems, such as resells or private information collected without consent, for example. Self managed data, or SKDs, are agents endowed with data, capabilities and goals to achieve. They are free to behave as they wish and try to accomplish their goals as efficiently as possible. They use learning algorithms to improve their decision making and learn new capabilities and services. We introduced how SKDs could be developed and provided some insight on useful capabilities.

In 10 we investigated to show a way to define autonomous data as well as some challenges associated with their specificities. A first version of a SKYDATA Environment is being developed using JADE and JASON to implement SKDs using the Beliefs-Desires-Intentions model with AgentSpeak.

Modeling and simulation are crucial in high-performance computing (HPC), with numerous frameworks developed for distributed computing infrastructures and their applications. Despite node-level simulation of shared-memory systems and task-based parallel applications, existing works overlook non-uniform memory access (NUMA) effects, a critical characteristic of current HPC platforms. In this work, we introduce a modeling for complex NUMA architectures and enhance a simulator for dependency-based task-parallel applications. This facilitates experiments with varied data locality models: we refine a communication-oriented model leveraging topology information for data transfers, and devise a more intricate model incorporating a cache mechanism for last-level cache data storage. Dense linear algebra test cases are used to validate both models, demonstrating that our simulator reliably predicts execution time with minimal relative error. This work 6 was one of the result obtained during the PhD of Idriss Douadi co-advised by Samuel Thibault (Prof. Univ. Bordeaux, EPI STORM) and Thierry Gautier.

Many-core and heterogeneous architectures now require programmers to compose multiple asynchronous programming model to fully exploit hardware capabilities. As a shared-memory parallel programming model, OpenMP has the responsibility of orchestrating the suspension and progression of asynchronous operations occurring on a compute node, such as MPI communications or CUDA/HIP streams. Yet, specifications only come with the task detach(event) API to suspend tasks until an asynchronous operation is completed, which presents a few drawbacks. In this paper, we introduce the design and implementation of an extension on the taskwait construct to suspend a task until an asynchronous event completion. It aims to reduce runtime costs induced by the current solution, and to provide a standard API to automate portable task suspension solutions. The results show twice less overheads compared to the existing task detach clause. This work 11 was one of the result obtained during the PhD of Romain Peirera co-advised by Adrien Roussel (CEA), Patrick Carribaut (CEA) and Thierry Gautier.

The paper 13 introduces a new approach to handle implicit parallelism in library functions. If the library already utilizes a thirdparty programming model like OpenMP, it may run in parallel. Otherwise, if the library remains sequential, OpenMP directives in client code cannot be used for direct parallelization. To express implicit parallelism and, in the meanwhile, dynamically adjust the parallel degree of a task when its starts, we propose to use moldable tasks. We handle this by introducing a new construct called taskmoldable that generates multiples tasks from a single function call and an iteration space. For the Lapack Cholesky factorization algorithm, our taskmoldable directive allows simple code annotation to express parallelism between tiles and improves programmability. Performance results on a beamforming application indicates that our moldable implementation is slightly faster by 5% in mean, than a parallel execution achieved with Intel MKL. This work is one of the result of the PhD of Pierre-Etienne Polet co-advised by Ramy Fantar(Thales) and Thierry Gautier

The architecture of supercomputers is evolving to expose massive parallelism. MPI and OpenMP are widely used in application codes on the largest supercomputers in the world. The community primarily focused on composing MPI with OpenMP before its version 3.0 introduced task-based programming. Recent advances in OpenMP task model and its interoperability with MPI enabled fine model composition and seamless support for asynchrony. Yet, OpenMP tasking overheads limit the gain of task-based applications over their historical loop parallelization (parallel for construct). This paper identifies the OpenMP task dependency graph discovery speed as a limiting factor in the performance of task-based applications. We study its impact on intra and inter-node performances over two benchmarks (Cholesky, HPCG) and a proxy-application (LULESH). We evaluate the performance impacts of several discovery optimizations, and introduce a persistent task dependency graph reducing overheads by a factor up to 15 at run-time. We measure 2x speedup over parallel for versions weak scaled to 16K cores, due to improved cache memory use and communication overlap, enabled by task refinement and depth-first scheduling. This work 12 was one of the result obtained during the PhD of Romain Peirera co-advised by Adrien Roussel (CEA), Patrick Carribaut (CEA) and Thierry Gautier.

Heterogeneous supercomputers with GPUs are one of the best candidates to build Exascale machines. However, porting scientific applications with millions of lines of code lines is challenging. Data transfers/locality and exposing enough parallelism determine the maximum achievable performance on such systems. Thus porting efforts impose developers to rewrite parts of the application which is tedious and time-consuming and does not guarantee performances in all the cases. Being able to detect which parts can be expected to deliver performance gains on GPUs is therefore a major asset for developers. Moreover, task parallel programming model is a promising alternative to expose enough parallelism while allowing asynchronous execution between CPU and GPU. OpenMP 4.5 introduces the « target » directive to offload computation on GPU in a portable way. Target constructions are considered as explicit OpenMP task in the same way as for CPU but executed on GPU. In this work 15, we propose a methodology to detect the most profitable loops of an application that can be ported on GPU. While we have applied the detection part on several mini applications (LULESH, miniFE, XSBench and Quicksilver), we experimented the full methodology on LULESH through MPI+OpenMP task programming model with target directives. It relies on runtime modifications to enable overlapping of data transfers and kernel execution through tasks. This work has been integrated into the MPC framework, and has been validated on distributed heterogeneous system.

We defined here some guidelines for critical applications to reduce the arithmetic numerical issue and we provide additional guidelines dedicated to Cloud platform. Many architecture was evaluted (OS native, Virtual Machine or container architecture). With some care, porting numerical software from a micro-controller to the Cloud, or directly writing applications to the Cloud, can be done provided that some recommendations listed below be followed. It is in fact not more difficult than porting software from a micro-controller to any general-purpose processor. The result of this research is an internal unpublished report dedicated to a private partner.