Program 2022

The widespread adoption of continuously connected smartphones and tablets drove the proliferation of mobile applications, among which many use location to provide a geolocated service. The usefulness of these services is no more to be demonstrated; getting directions to work in the morning, leaving a check-in at a restaurant at noon and checking next day’s weather in the evening is possible from any mobile device embedding a GPS chip. In these applications, locations are sent to a server often hosted on untrusted cloud platforms, which uses them to provide personalized answers. However, nothing prevents these platforms from gathering, analyzing and possibly sharing the collected information. This opens the door for many threats, as location information allows to infer sensitive information about users, among which one’s home, workplace or even religious/political preferences. For this reason, many schemes have been proposed these last years to enhance location privacy while still allowing people to enjoy geolocated services. During this presentation, I will present the latest advances in location privacy attacks and protection mechanisms and give some insights on open challenges and under-explored questions.

Sonia Ben Mokhtar is a CNRS research director at the LIRIS laboratory (UMR 5205) and the head of the distributed systems and information retrieval group (DRIM). She received her PhD in 2007 from Université Pierre et Marie Curie before spending two years at University College London (UK). Her research focuses on the design of resilient and privacy-preserving distributed systems. Sonia has co-authored 70+ papers in peer-reviewed conferences and journals and has served on the editorial board of IEEE Transactions on Dependable and Secure Computing and co-chaired major conferences in the field of distributed systems (e.g., ACM Middleware, IEEE DSN). Sonia has served as chair of ACM SIGOPS France and is currently the vice-chair of GDR RSD a national academic network of researchers in distributed systems and networks.

Constraint programming is a field of artificial intelligence that deals with automatic problem solving. It is a declarative paradigm in which the user describes a problem through rules that its solutions must respect: constraints. In this presentation, we will see that constraints are much more than modelling tools and participate in problem solving. In a didactic way, we will explore the functioning of a constraint solver and see why CP can become part of your decision-making toolbox.

Chales Prudhomme is an assistant professor at IMT Atlantique and researcher in the TASC team at LS2N (Nantes, France). He obtained his PhD from Ecole des Mines de Nantes in 2014 and then worked as a research engineer. His research focuses on the design and improvement of constraint solvers, from tools to applications. For the past 15 years, he has been the lead developer of choco-solver, an open-source constraint programming library.

The complexity of parallel architectures continues to increase. Multicore processors, non-uniform memory access (NUMA), and deep hierarchies of shared or private caches raised the need to carefully place computing tasks and data buffers together. Heterogeneity added yet another layer of difficulty by requiring to identify tasks that can be efficiently processed by GPUs or CPUs. Heterogeneity is now visible in the memory subsystems where different kinds of memory (DRAM, HBM, non-volatile) may co-exist, and it is even coming inside CPUs (energy-efficient small cores versus power-hungry big cores). Exploiting the full performance of these modern platforms implies a deep knowledge of the hardware. However application developers cannot be hardware experts, and portability is not compatible with manual tuning for a specific platforms. In this presentation, we will discuss how to model these complex platforms in a portable manner and how application developers should exhibit their needs so that the software needs can be matched with the hardware abilities for better performance thanks to better placement.

Brice Goglin is an Inria research director and member of the LaBRI laboratory and TADaaM Inria team in Bordeaux. He earnt his PhD in Lyon in 2005 and spent one year at Myricom (Tennessee) working on high-speed networks for high-performance computing (HPC). He joint Inria in 2006 and worked on the management of multicore and NUMA architectures. This involved task and data placement, as well as adapting communication schemes to the placement. His team research on modeling parallel computing platforms led to the development of the hwloc software (Hardware Locality) which is now used by the vast majority of parallel libraries in HPC for abstracting the topology of hardware resources (cores, memory, accelerators, etc).

The scale and intensity of today's economic exchanges require modern web services to be always-on and responsive. One of the key concerns when building these services is the durability and availability of their data. Data must remain accessible despite network, machine, or even data center-scale temporary (or permanent) outages. This talk explores some of the recent techniques to achieve such guarantees.

First, we focus on geo-replicated state machines. Geo-replication places several copies of a logical data item across different data centers to improve access locality and fault-tolerance. State-machine replication (SMR) ensures that these copies stay in sync. Recent advances in SMR focus on leaderless protocols that sidestep the availability and performance limitations of traditional Paxos-based solutions. We detail several leaderless protocols and compare their pros and cons under different applicative workloads.

The second part of this talk focuses on non-volatile main memory (NVMM). NVMM is a new tier in the memory hierarchy that offers jointly the durability of spinning disks, near-DRAM speed and byte addressability. We present J-NVM, a fully-fledged interface to use NVMM in the Java language. Internally, J-NVM relies on proxy objects that intermediate direct off-heap access to NVMM. We present the internal of J-NVM and how to use it in the context of a modern distributed data store.

Pierre Sutra is Associate Professor at Télécom SudParis and a committer of the Apache Software Foundation. His work investigates the theory and practice of distributed systems, with applications to big data stores, transactional systems and storage systems. Pierre has co-authored more than 40 research publications. Over the past two years, he has published at the following top journals and conferences: ACM TOSEM, Information Processing Letters, SOSP, EuroSys, DISC and Middleware.

Digital agriculture defines an agriculture and more generally food systems using information and communication technologies (ICT) as technologies for data capture (satellites, sensors, connected objects, smartphones…), transfer and storage (3G/4G/5G, terrestrial or satellite low rate networks, clouds), and processing (on-board or remote, artificial intelligence, …), at all scales (from farm to land). Its general goals encompass a better efficiency, a better animal welfare, more frugality (water, nitrates, chemical treatments, etc). In this context, we aim to deploy some wireless sensor network for accurate parcel-related crop monitoring. But such sensors must be able to send their data even from locations not covered by a traditional network such as cellular or wifi and at a low energy cost. This talk will present the research leading the LIRIMA AGRINET project, a joint project between Inria and Stellenbosch University aiming to address these issues.

Nathalie Mitton received the MSc and PhD. degrees in Computer Science from INSA Lyon in 2003 and 2006 respectively. She received her Habilitation à diriger des recherches (HDR) in 2011 from Université Lille 1. She is currently an Inria full researcher since 2006 and from 2012, she is the scientific head of the Inria FUN team which is focused on communication in wireless constrained networks. Her research interests focus on self-organization from PHY to routing for wireless constrained dynamic and mobile networks. She has been nominated as one of the 10 women stars in computer Science in 2020 by the IEEE Communication Society. She has published her research in more than 40 international revues and more than 120 international conferences. She is involved in the set up of the FIT IoT LAB platform (http://fit-equipex.fr/, https://www.iot-lab.info), the H2020 CyberSANE projects and in several program and organization committees such as Infocom (since 2019), PerCom (since 2018), DCOSS (since 2018), ICC (since 2015), Globecom (since 2017), etc. She also supervises several PhD students and engineers.

Data science relies on the processing of vast amount of data. While some this data may be “at rest” and stored on a distributed collection of disks, some may arrive in real time and must be processed as quickly as possible. Traditionally, the Big Data processing field distinguished between batch data processing (e.g., Apache Hadoop and Spark) and stream processing (e.g., Apache Storm). Latest-generation platforms such as Apache Flink reconcile the two models and allow processing large amounts of data at rest efficiently in conjunction with high-velocity stream data. Flink proposes different APIs allowing to create and operate dataflow operations over the data. In this lecture, we will position Flink and other major data processing engines for the cloud and their guiding principles, with a focus on the distributed systems aspects (performance, data management, fault tolerance). We will present how Flink stores local and global data efficiently and how it employs passive replication through snapshotting for fault tolerance and elastic reconfiguration. We will also present how Flink can be modified to support high-availability through active replication, i.e., operations that continue with low latency despite the failure of some of the components supporting the computation in a cluster.

Etienne Rivière is a professor of compter science at UCLouvain in Belgium. He leads the Cloud and Large-Scale Computing group (https://cloudlargescale-uclouvain.github.io/), par of UCLouvain’s Institute for Information and Communication Technologies, Electronics and Applied Mathematics (ICTEAM), and is vice-head of UCLouvain’s compter science department (INGI). Pr. Rivière obtained his PhD in 2007 from University of Rennes and Inria. He has been affiliated with NTNU, Norway and the University of Neuchâtel, Switzerland, prior to joining UCLouvain in 2017. His research interest are in computer systems, with a focus on scalability and performance, dependability, and security. Pr. Rivière and his team participate in several projects, some in close collaboration with the industry. For instance, the GEPICIAD project focuses on the management and orchestration of data processing frameworks such as Flink, in collaboration with a major R&D player in data science.

Trusted execution environments (TEEs) are increasingly popular hardware-based enabling technologies to improve the security guarantees of systems. In this keynote, I will first explain what are the main characteristics of TEEs for a few of the most popular ones (such as Intel SGX, AMD SEV and ARM TrustZone). Then, I will present our recent work to support the execution of WebAssembly (WASM), a modern and fast binary instruction format, into SGX enclaves. Twine, our shielded WASM runtime, allows for the shielded execution of unmodified WASM applications into SGX enclaves. I will present some of the extensions done to WASI (the WASM System Interface) as well as some evaluation results using real-world database systems (ie, SQLite).

Valerio Schiavoni received his M.Sc. and Ph.D. degrees in Computer Science from Roma Tre University (Italy) and from the University of Neuchâtel (Switzerland), respectively. Until 2017 he was the scientific coordinator of the Centre of Competence for Complex Systems and Big Data (CC-CSBD) at University of Neuchâtel. Until 2021, he coordinated CUSO (Conférence Universitaire Suisse Occidentale) for the Computer Science programs. He co-founded one start-up, and more recently co-founded the ARM HPC User Group (AHUG). Since 2018, he has been a Lecturer (Maître-Assistant) at the University of Neuchâtel. He is interested in systems broadly conceived.