Performance Characterization of Low-Level Building Blocks

Although Cloud Computing has enabled the consolidation of services and applications into a subset of servers, current operating system mechanisms do not provide appropriate abstractions to prevent (or at least control) the performance degradation that occurs when several workloads compete for the same resources  95. Keeping in mind that server density is going to increase with physical machines composed of more and more cores and that applications will be more and more data intensive, it is mandatory to identify interferences that appear at a low level on each dimension (compute, memory, network, and storage) and propose counter-measures. In particular, previous studies  95, 54 on pros and cons of current technologies – virtual machines (VMs)  71, 79, containers and microservices – which are used to consolidate applications on the same server, should be extended: In addition to evaluating the performance we can expect from each of these technologies on a single node, it is important to investigate interferences that may result from cross-layer and remote communications  96. We will consider in particular all interactions related to geo-distributed systems mechanisms/services that are mandatory to operate and use geo-distributed ICT infrastructures.

Geo-Distributed System Mechanisms

Although several studies have been highlighting the advantages of geo-distributed ICT infrastructures in various domains (see Section 4), progress on how to operate and use such infrastructures is marginal. Current solutions  25  26 are rather close to the initial Cisco Fog Computing proposal that only allows running domain-specific applications on edge resources and centralized Cloud platforms  33 (in other words, these solutions do not allow running stateful workloads in isolated environments such as containers or VMs). More recently, solutions leveraging the idea of federating VIMs (as the aforementioned ETSI MEC proposal  84) have been proposed. ONAP  58, an industry-driven solution, enables the orchestration and automation of virtual network functions across distinct VIMs. From the academic side, FogBow  36 aims to support federations of Infrastructure-as-a-Service (IaaS) providers. Finally, NIST initiated a collaborative effort with IEEE to advance Federated Cloud platforms through the development of a conceptual architecture and a vocabulary2. Although all these projects provide valuable contributions, they face the aforementioned orchestration limitations (i.e., they do not manage decisions taken in each VIM). Moreover, they all have been designed by only considering the developer/user's perspective. They provide abstractions to manage the life cycle of geo-distributed applications, but do not deliver means to administer the physical resources.

To cope with specifics of Wide-Area networks while delivering most features that made Cloud Computing solutions successful also at the edge, our community should first identify limitations/drawbacks of current resource management system mechanisms with respect to the Fog/Edge requirements and propose revisions when needed  64, 77.

To achieve this aim, STACK members propose to conduct first a series of studies aiming at understanding the software architecture and footprint of major services that are mandatory for operating and using Fog/Edge infrastructures (storage backends, monitoring services, deployment/reconfiguration mechanisms, etc.). Leveraging these studies, we will investigate how these services should be deployed in order to deal with resources constraints, performance variability, and network split brains. We will rely on contributions that have been accomplished in distributed algorithms and self-* approach for the last decade. In the short and medium term, we plan to evaluate the relevance of NewSQL systems  46 to store internal states of distributed system mechanisms in an edge context, and extend our proposals on new storage backends such as key/value stores  45, 90, and burst buffers  97. We also plan to conduct new investigations on data-stream frameworks for Fog and Edge infrastructures  40. These initial contributions should enable us to identify general rules to deliver other advanced system mechanisms that will be mandatory at the higher levels in particular for the deployment and reconfiguration manager in charge of orchestrating all resources.

Capacity Planning and Placement Strategies

An objective shared by users and providers of ICT infrastructures is to limit as much as possible the operational costs while providing the expected and requested quality of service (QoS). To optimize this cost while meeting QoS requirements, data and applications have to be placed in the best possible way onto physical resources according to data sources, data types (stream, graphs), application constraints (real-time requirements) and objective functions. Furthermore, the placement of applications must evolve through time to cope with the fluctuations in terms of application resource needs as well as the physical events that occur at the infrastructure level (resource creation/removals, hardware failures, etc.). This placement problem, a.k.a. the deployment and reconfiguration challenge as it will be described in Section 3.3, can be modeled in many different ways, most of the time by multi-dimensional and multi-objective bin-packing problems or by scheduling problems which are known to be difficult to solve. Many studies have been performed, for example, to optimize the placement of virtual machines onto ICT infrastructures  73. STACK will inherit the knowledge acquired through previous activities in this domain, particularly its use of constraint programming strategies in autonomic managers  69, 68, relying on MAPE (monitor, analyze, plan, and execute) control loops. While constraint programming approaches are known to hardly scale, they enable the composition of various constraints without requiring to change heuristic algorithms each time a new constraint has to be considered  67. We believe it is a strong advantage to deal with the diversity brought by geo-distributed ICT infrastructures. Moreover, we have shown in previous work that decentralized approaches can tackle the scalability issue while delivering placement decisions good enough and sometimes close to the optimal  83.

Leveraging this expertise, we propose, first, to identify new constraints raised by massively geo-distributed infrastructures (e.g., data locality, energy, security, reliability and the heterogeneity and mobility of the underlying infrastructure). Based on this preliminary study, we will explore new placement strategies not only for computation sandboxes but for data (location, replication, streams, etc.) in order to benefit from the geo-distribution of resources and meet the required QoS. These investigations should lead to collaborations with operational research and optimization groups such as TASC, another research group from IMT Atlantique.

Second, we will leverage contributions made on the previous axis “Performance Characterization of Low-Level Building Blocks” to determine how the deployment of the different units (software components and data sets) should be executed in order to reduce as much as possible the time to reconfigure the system (i.e., the Execution phase in the control loop). In some recent work  79, we have shown that the provisioning of a new virtual machine should be done carefully to mitigate boot penalties. More generally, proposing an efficient action plan for the Execution phase will be a major point as Wide-Area-Network specifics may lead to significant delays, in particular when the amount of data to be manipulated is important.

Finally, we will investigate new approaches to decentralize the placement process while considering the geo-distributed context. Among the different challenges to address, we will study how a combination of autonomic managers, at both the infrastructure and application levels  53, could be proposed in a decentralized manner. Our first idea is to geo-distribute a fleet of small control loops over the whole infrastructure. By improving the locality of data collection and weakening combinatorics, these loops would allow the system to address responsiveness and quality expectations.

Programming Models and Languages Extensions

The current landscape of programming support for cloud applications is fragmented. This fragmentation is based on apparently different needs for various kinds of applications, in particular, web-based, computation-based, focusing on the organization of the computation, and data-based applications, within the last case a quite strong dichotomy between applications considering data as sets or relations, close to traditional database applications and applications considering data as real-time streams. This has led to various programming models, in a loose sense, including for instance microservices, graph processing, dataflows, streams, etc. These programming models have mostly been offered to the application programmer in the guise of frameworks, each offering subtle variants of the programming models with various implementation decisions favoring particular application and infrastructure settings. Whereas most frameworks are dedicated to a given programming model, e.g., basic Pregel  78, Hive  91, Hadoop  92, some of them are more general-purpose through the provision of several programming models, e.g., Flink  39 and Spark  75. Finally, some dedicated language support has been considered for some models (e.g., the language SPL underlying IBM Streams  70) as well as core languages and calculi (e.g.,   35, 88).

This situation raises a number of challenges on its own, related to a better structuring of the landscape. It is necessary to better understand the various programming models and their possible relations, with the aim of facilitating, if not their complete integration, at least their composition, at the conceptual level but also with respect to their implementations, as specific languages and frameworks.

Switching to massively geo-distributed infrastructures adds to these challenges by leading to a new range of applications (e.g., smart-* applications) that, by nature, require mixing these various programming models, together with a much more dynamic management of their runtime.

In this context, STACK would like to explore two directions:

• First, we propose to contribute to generic programming models and languages to address composability of different programming models  48. For example, providing a generic stream data processing model that can operate under both data stream  39 and operation stream  98 modes, thus streams can be processed in micro batches to favour high throughput or record by record to sustain low latency. Software engineering properties such as separation of concerns and composition should help address such challenges  28, 89. They should also facilitate the software deployment and reconfiguration challenges discussed below.

• Second, we plan to revise relevant programming models, the associated specific languages, and their implementation according to the massive geo-distribution of the underlying infrastructure, the data sources, and application end-users. For example, although SPL is extensible and distributed, it has been designed to run on multi-cores and clusters  70. It does not provide the level of dynamicity required by geo-distributed applications (e.g., to handle topology changes, loss of connectivity at the edge, etc.).

  Moreover, as more network data transfers will happen within a massively geo-distributed infrastructure, correctness of data transfers should be guaranteed. This has potential impact from the programming models to their implementations.

Deployment and Reconfiguration Challenges

The second research direction deals with the complexity of deploying distributed software (whatever the layer, application, middleware or system) onto an underlying infrastructure. As both the deployed pieces of software and the infrastructures addressed by STACK are large, massively distributed, heterogeneous and highly dynamic, the deployment process cannot be handled manually by developers or administrators. Furthermore, and as already mentioned in Section 3.2, the initial deployment of some distributed software will evolve through time because of the dynamicity of both the deployed software and the underlying infrastructures. When considering reconfiguration, which encompasses deployment as a specific case, the problem becomes more difficult for two main reasons: (1) the current state of both the deployed software and the infrastructure has to be taken into account when deciding on a reconfiguration plan, (2) as the software is already running the reconfiguration should minimize disruption time, while avoiding inconsistencies  76, 81. Many deployment tools have been proposed both in academia and industry  50. For example, Ansible, Chef and Puppet are very well-known generic tools to automate the deployment process through a set of batch instructions organized in groups (e.g., playbooks in Ansible). Some tools are specific to a given environment, like Kolla to deploy OpenStack, or the embedded deployment manager within Spark. Few reconfiguration capabilities are available in production tools such as scaling and restart after a fault, for example in Kubernetes and Juju charms. Academia has contributed to generic deployment and reconfiguration models. Most of these contributions are component-based. Component models divide a distributed software as a set of component instances (or modules) and their assembly, where components are connected through well defined interfaces  89. Thus, modeling the reconfiguration process consists in describing the life cycle of different components and their interactions. Most component-based approaches offer a fixed life cycle, i.e., identical for any component  55. Two main contributions are able to customize life cycles, Fractal  38, 31 and its evolutions  28, 29, 52, and Aeolus  47. In Fractal, the control part of a component (e.g., its life cycle) is modeled itself as a component assembly that is highly flexible. Aeolus, on the other hand, offers a finer control on both the evolution and the synchronization of the deployment process by modeling each component life cycle with a finite state machine.

A reconfiguration raises at least five questions, all of them are correlated: (1) why software has to be reconfigured? (monitoring, modeling and analysis) (2) what should be reconfigured? (software modeling and analysis), (3) how should it be reconfigured? (software modeling and planning decisions), (4) where should it be reconfigured? (infrastructure modeling and planning decisions), and (5) when to reconfigure it? (scheduling algorithms). STACK will contribute to all aspects of a reconfiguration process as described above. However, according to the expertise of STACK members, we will focus mainly on the three first questions: why, what and how, leaving questions where and when to collaborations with operational research and optimization teams.

First of all, we would like to investigate why software has to be reconfigured? Many reasons could be mentioned, such as hardware or software fault tolerance, mobile users, dynamicity of software services, etc. All those reasons are related somehow to the Quality of Service (QoS) or the Service Level Agreement (SLA) between the user and the Cloud provider. We first would like to explore the specificities of QoS and SLAs in the case of massively geo-distributed ICT environments  85. By being able to formalize this question, analyzing the requirement of a reconfiguration will be facilitated.

Second, we think that four important properties should be enhanced when deploying and reconfiguring models in massively geo-distributed ICT environments. First, as low-latency applications and systems will be subject to deployment and reconfiguration, the performance and the ability to scale are important. Second, as many different kinds of deployments and reconfigurations will concurrently hold within the infrastructure, processes have to be reliable, which is facilitated by a fine-grained control of the process. Finally, as many different software elements will be subject to deployment and reconfiguration, common generic models and engines for deployment and reconfiguration should be designed  37. For these reasons, we intend to go beyond Aeolus by: first, leveraging the expression of parallelism within the deployment process, which should lead to better performance; second, improving the separation of concerns between the component developer and the reconfiguration developer; third, enhancing the possibility to perform concurrent and decentralized reconfigurations.

Research challenges relative to programming support have been presented above. Many of these challenges are related, in different manners, to the resource management level of STACK or to crosscutting challenges, i.e., energy and security. First, one can notice that any programming model or deployment and reconfiguration implementation should be based on mechanisms related to resource management challenges. For this reason, all challenges addressed within this section are linked with lower level building blocks presented in Section 3.2. Second, as detailed above, deployment and reconfiguration address at least five questions. The question what? is naturally related to programming support. However, questions why, how?, where? and when? are also related to Section 3.2, for example, to monitoring and capacity planning. Moreover, regarding the deployment and reconfiguration challenges, one can note that the same goals recursively happen when deploying the control building blocks themselves (bootstrap issue). This comforts the need to design generic deployment and reconfiguration models and frameworks. These low-level models should then be used as back-ends to higher-level solutions. Finally, as energy and security are crosscutting themes within the STACK project, many additional energy and security considerations could be added to the above challenges. For example, our deployment and reconfiguration frameworks and solutions could be used to guarantee the deployment of end-to-end security policies or to answer specific energy constraints  66 as detailed in the next section.

7.1.1 MAD

• Name:
  Madeus Application Deployer
• Keywords:
  Automatic deployment, Distributed Software, Component models, Cloud computing
• Scientific Description:
  MAD is a Python implementation of the Madeus deployment model for multi-component distributed software. Precisely, it allows to: 1. describe the deployment process and the dependencies of distributed software components in accordance with the Madeus model, 2. describe an assembly of components, resulting in a functional distributed software, 3. automatically deploy the component assembly of distributed software following the operational semantics of Madeus.
• Functional Description:
  MAD is a Python implementation of the Madeus deployment model for multi-component distributed software. Precisely, it allows to: 1. describe the deployment process and the dependencies of distributed software components in accordance with the Madeus model, 2. describe an assembly of components, resulting in a functional distributed software, 3. automatically deploy the component assembly of distributed software following the operational semantics of Madeus.
• Release Contributions:
  Initial submission with basic functionalities of MAD
• News of the Year:
  Operational prototype.
• Publications:
  hal-01858150, hal-01897803
• Contact:
  Hélène Coullon
• Participants:
  Christian Perez, Dimitri Pertin, Hélène Coullon, Maverick Chardet
• Partners:
  IMT Atlantique, LS2N, LIP

7.1.2 Nitro

• Keywords:
  Cloud storage, Virtual Machine Image, Geo-distribution
• Scientific Description:

  Nitro is a storage system that is designed to work in geo-distributed cloud environments (i.e., over WAN) to efficiently manage Virtual Machine Images (VMIs).

  Nitro employs fixed-size deduplication to store VMIs. This technique contributes to minimizing the network cost. Also, Nitro incorporates a network-aware scheduling algorithm (based on max flow algorithm) to determine which chunks should be pulled from which site in order to reconstruct the corresponding image on the destination site, with minimal (provisioning) time.

• Functional Description:
  Geo-distributed Storage System to optimize Images (VM, containers, ...) management, in terms of cost and time, in geographically distributed cloud environment (i.e. data centers are connected over WAN).
• URL:
  https://­gitlab.­inria.­fr/­jdarrous/­nitro
• Authors:
  Jad Darrous, Shadi Ibrahim, Christian Perez
• Contact:
  Shadi Ibrahim

7.1.3 VMPlaceS

• Keywords:
  Simulation, Virtualization, Scheduling
• Functional Description:

  VMPlaces is a dedicated framework to evaluate and compare VM placement algorithms. This framework is composed of two major components: the injector and the VM placement algorithm. The injector is the generic part of the framework (i.e. the one you can directly use) while the VM placement algorithm is the part you want to study (or compare with available algorithms). Currently, the VMPlaceS is released with three algorithms:

  Entropy, a centralized approach using a constraint programming approach to solve the placement/reconfiguration VM problem

  Snooze, a hierarchical approach where each manager of a group invokes Entropy to solve the placement/reconfiguration VM problem. Note that in the original implementation of Snooze, it is using a specific heuristic to solve the placement/reconfiguration VM problem. As the sake of simplicity, we have simply reused the entropy scheduling code.

  DVMS, a distributed approach that dynamically partitions the system and invokes Entropy on each partition.

• URL:
  http://­beyondtheclouds.­github.­io/­VMPlaceS/
• Contact:
  Adrien Lebre
• Participants:
  Adrien Lebre, Jonathan Pastor, Mario Südholt

7.1.4 ENOS

• Name:
  Experimental eNvironment for OpenStack
• Keywords:
  OpenStack, Experimentation, Reproducibility
• Functional Description:

  Enos workflow :

  A typical experiment using Enos is the sequence of several phases: - enos up : Enos will read the configuration file, get machines from the resource provider and will prepare the next phase - enos os : Enos will deploy OpenStack on the machines. This phase rely highly on Kolla deployment. - enos init-os : Enos will bootstrap the OpenStack installation (default quotas, security rules, ...) - enos bench : Enos will run a list of benchmarks. Enos support Rally and Shaker benchmarks. - enos backup : Enos will backup metrics gathered, logs and configuration files from the experiment.

• URL:
  https://­github.­com/­BeyondTheClouds/­enos
• Contact:
  Adrien Lebre
• Partner:
  Orange Labs

7.1.5 EnOSlib

• Name:
  EnOSlib is a library to help you with your experiments
• Keywords:
  Distributed Applications, Distributed systems, Evaluation, Grid Computing, Cloud computing, Experimentation, Reproducibility, Linux, Virtualization
• Functional Description:

  EnOSlib is a library to help you with your distributed application experiments. The main parts of your experiment logic is made reusable by the following EnOSlib building blocks:

  - Reusable infrastructure configuration: The provider abstraction allows you to run your experiment on different environments (locally with Vagrant, Grid’5000, Chameleon and more) - Reusable software provisioning: In order to configure your nodes, EnOSlib exposes different APIs with different level of expressivity - Reusable experiment facilities: Tasks help you to organize your experimentation workflow.

  EnOSlib is designed for experimentation purpose: benchmark in a controlled environment, academic validation …

• URL:
  https://­discovery.­gitlabpages.­inria.­fr/­enoslib/
• Publications:
  hal-01664515, hal-01689726
• Contact:
  Mathieu Simonin

7.1.6 Concerto

• Name:
  Concerto
• Keywords:
  Reconfiguration, Distributed Software, Component models, Dynamic software architecture
• Scientific Description:
  Concerto is a reconfiguration model which allows to describe distributed software as an evolving assembly of components.
• Functional Description:
  Concerto is an implementation of the formal model Concerto written in Python. Concerto allows to : 1. describe the life-cycle and the dependencies of software components, 2. describe a components assembly that forms the overall life-cycle of a distributed software, 3. automatically reconfigure a Concerto assembly of components by using a set of reconfiguration instructions as well as a formal operational semantics.
• News of the Year:
  In 2020, we added the ability to read from and write to non-data provide ports. We also updated the Madeus wrapper to the new version with only USE and PROVIDE ports and no groups.
• URL:
  https://­gitlab.­inria.­fr/­VeRDi-project/­concerto
• Publications:
  hal-03103714, hal-02535077, hal-01897803
• Contact:
  Hélène Coullon
• Partners:
  IMT Atlantique, LS2N, LIP

7.2.1 OpenStack

OpenStack is the de facto open-source management system to operate and use Cloud Computing infrastructures. Started in 2012, the OpenStack foundation gathers 500 organizations including groups such as Intel, AT&T, RedHat, etc. The software platform relies on tens of services with a 6-month development cycle. It is composed of more than 2 millions of lines of code, mainly in Python, just for the core services. While these aspects make the whole ecosystem quite swift, they are also good signs of maturity of this community.

We created and animated between 2016 and 2018 the Fog/Edge/Massively Distributed (FEMDC) Special Interest Group and have been contributing to the Performance working group since 2015. The former investigates how OpenStack can address Fog/Edge Computing use cases whereas the latter addresses scalability, reactivity and high-availability challenges. In addition to releasing white papers and guidelines 56, the major result from the academic view point is the aforementioned EnOS solution, a holistic framework to conduct performance evaluations of OpenStack (control and data plane). In May 2018, the FEMDC SiG turned into a larger group under the control of the OpenStack foundation. This group gathers large companies such as Verizon, ATT, etc. Although our involvment has been less important in 2020, our participation is still signficant. For instance, we co-signed the second white paper delivered by the edge working group in 2020  57.

7.2.2 Grid'5000

Grid'5000 is a large-scale and versatile testbed for experiment-driven research in all areas of computer science, with a focus on parallel and distributed computing including Cloud, HPC and Big Data. It provides access to a large amount of resources: 12000 cores, 800 compute-nodes grouped in homogeneous clusters, and featuring various technologies (GPU, SSD, NVMe, 10G and 25G Ethernet, Infiniband, Omni-Path) and advanced monitoring and measurement features for traces collection of networking and power consumption, providing a deep understanding of experiments. It is highly reconfigurable and controllable. STACK members are strongly involved into the management and the supervision of the testbed, notably through the steering committee or the SeDuCe testbed described hereafter.

7.2.3 SeDuCe

The SeDuCe Project aims to deliver a research testbed dedicated to holistic research studies on energetical aspects of datacenters. Part of the Grid'5000 Nantes' site, this infrastructure is composed of probes that measure the power consumption of each server, each switch and each cooling system, and also measure the temperature at the front and the back of each servers. These sensors enable reasearch to cover a full spectrum of the energetical aspect of datacenters, such as cooling and power consumption depending of experimental conditions.

The testbed is connected to renewable energy sources (solar panels). This “green” datacenter enables researchers to perform real experiment-driven studies on fields such as temperature based scheduling or “green” aware software (i.e., software that take into account renewable energies and weather conditions).

7.2.4 PiSeDuCe

In 2021, we consolidated the development of PiSeDuCe, a deployment and reservation system for Edge Computing infrastructures composed of multiple Raspberry Pi Cluster started in 2020. Typically, a cluster of 8 Raspberry Pi costs less than 900 euros and only needs an electrical outlet and a wifi connection for its installation and configuration. Funded by the CNRS through the Kabuto project, and in connection with the SILECS initiative, we have extended PiSeduce to propose a device to cloud deployment system (from devices on Fit IoTLab to servers in Grid’5000).

7.2.5 SILECS/SLICES

STACK Members are involved in the definition and bootstrap of the SILECS infrastructure. This infrastructure can be seen as a merge of the Grid'5000 and FIT testbeds with the goal of providing a common platform for experimental computer Science (Next Generation Internet, Internet of things, clouds, HPC, big data, etc.). In 2021, STACK contributions are mainly related to the European initiatives relative to SILECS. In particular, members have taken part to the SLICES-DS project (Design Study) as well as the submission of SLICES-PP (Preparatory Phase) of the SLICES-RI action .

Resource management systems/middleware for Edge infrastructures:

The two first contributions have been conducted within the framework of the Inria/Orange Laboratory. More precisely, we have investigated how connectivity among several resource managers in charge of operating, each one a subset of the infrastructure, could be established. In 3, we have surveyed and analyzed the characteristics and limitations of existing technologies in the Software Defined Network field that could be used to provide the intersite connectivity feature. We have also introduced how the network is addressed in the Kubernetes ecosystem, and have analyzed its use in the proposed context. We have concluded by providing a discussion about some research directions in the field of SDN applied to distributed Cloud-Edge infrastructures’ management. Leveraging this survey, we propose the DIMINET solution 22, a service in charge of providing on-demand connectivity for multiple sites. DIMINET leverages a logically and physically distributed architecture where instances collaborate on-demand and with minimal traffic exchange to provide inter-site connectivity management. The lessons learned during this study allows us to propose the premises of a generalization in order to be able to distribute in a non-intrusive manner any service in a DCI.

In 16, we have extended the aforementioned study on the vanilla Kubernetes framework with in-vivo experiments. Cloud Computing has highlighted the importance of container orchestration such as Kubernetes to manage the life-cycle of distributed applications. With the advent of Edge Computing, DevOps expect to find the features of containers in the cloud, also at the edge. However, similarly to systems such as OpenStack, orchestration solutions have not been designed to deal with geo-distribution aspects such as latency, intermittent networks, etc. In other words, it is unclear whether they could be directly used on top of massively distributed edge infrastructures without revision. To answer this question, we have conducted an evaluation of Kubernetes in a WANWide context leveraging the Grid'5000 testbed. More precisely, we have discussed results we obtained during an experimental campaign that aimed at analyzing the impact of WAN links on its behaviour. While multiple initiatives investigate how Kubernetes could be revised to deal with distribution aspects, this work, to the best of our knowledge, is the first one to rigorously evaluate whether the Kubernetes vanilla code could be directly used without any change.

In 24, we investigated how peers in a geo-distributed infrastructure could maintain an index of relevant replicas. Although the holy grail of storing and manipulating data in Edge infrastructures is yet to be found, state-of-the-art approaches demonstrated the relevance of replication strategies that bring content closer to consumers: The latter enjoy better response time while the volume of data passing through the network decreases overall. Unfortunately, locating the closest replica of a specific content requires indexing every live replica along with its location. Relying on remote services enters in contradiction with the properties of Edge infrastructures as locating replicas may effectively take more time than actually downloading content. At the opposite end, maintaining such an index at every node would prove overly costly in terms of memory and traffic, especially since nodes can create and destroy replicas at any time. We proposed to abstract this content indexing challenge as a distributed partitioning problem: every node only indexes its closest replica, and connected nodes with a similar index compose a partition. Our decentralized implementation AS-cast is (i) efficient, for it uses partitions to lock down the traffic generated by its operations to relevant nodes, yet it (ii) guarantees that every node eventually acknowledges its partition despite concurrent operations. Our complexity analysis supported by simulations shows that AS-cast scales well in terms of generated traffic and termination time. As such, AS-cast can constitute a new building block for geo-distributed services.

IoT resource management:

 As aforementioned, the team has initiated new activities covering the Cloud–IoT continuum.

In 10, we investigated the challenges and limitations in using a distributed network resource allocation mechanism in the wireless Industrial IoT context. The wireless communication protocol IEEE Std 802.15.4-2015 Time Slotted Channel Hopping (TSCH) is the de facto Medium Access Control (MAC) mechanism for industrial applications. It renders communications more resilient to interference by spreading them over the time (time-slotted) and the frequency (channel-hopping) domains. The 6TiSCH architecture bases itself on this new MAC layer to enable high reliability communication in Wireless Sensor Networks (WSNs). In particular, it manages the construction of a distributed communication schedule that continuously adapts to changes in the network. Our investigation comprises first a thorough description of the 6TiSCH architecture, the 6TiSCH Operation Sublayer (6top), and the Minimal Scheduling Function (MSF). We then study its behavior and reactivity from low to high traffic rates by employing the Python-based 6TiSCH simulator. Our performance evaluation results demonstrate that the convergence pattern of MSF is the root cause of the majority of packet losses observed in the network. We also show that MSF is prone to over-provisioning of the network resources, especially in the case of varying traffic load. We propose a mathematical model to predict the convergence pattern of MSF. Finally we investigate the impact of varying parameters on the behavior of the scheduling function.

In 6, we investigate the trade-offs of computational, memory and network traffic resource consumption against reliability of different packet fragmentation and forwarding mechanisms in the wireless Industrial IoT context which places significant demands in terms of reliability on wireless connectivity. The IEEE Std 802.15.4-2015 standard was designed in response to these demands, and the IPv6 over Low power Wireless Personal Area Networks (6LoWPAN) adaptation layer was introduced to address (among other issues) its payload size limitations by performing packet compression and fragmentation. However, the standardised method does not cope well with low link-quality situations and, thus, we present the state-of-the-art Forward Error Correction (FEC) methods and introduce our own contribution, Network Coding FEC (NCFEC), to improve performance in these situations. We present and analyse the existing methods as well as our own theoretically, and we then implement them and perform an experimental evaluation using the 6TiSCH simulator. The simulation results demonstrate that when high reliability is required and only low quality links are available, NCFEC performs best, with a trade-off between additional network and computational overhead. In situations where the link quality can be guaranteed to be higher, simpler solutions also start to be feasible, but with reduced adaptation flexibility.

Experiment-driven studies:

 The last contribution we achieved is related to our experiment-driven studies we have performed since 2016. Despite the importance of such activities in the distributed computing community, there has been little progress in helping researchers conduct their experiments. In most cases, we have to achieve tedious and time-consuming development and instrumentation activities to deal with the specifics of testbeds and the system under study. In order to relieve researchers of the burden of those efforts, we have developed ENOSLIB: a Python library that takes into account best experimentation practices and leverages modern toolkits on automatic deployment and configuration systems 2. ENOSLIB helps researchers not only in the process of developing their experimental artifacts, but also in running them over different infrastructures. To demonstrate the relevance of our library, we discuss three experimental engines built on top of ENOSLIB, and used to conduct empirical studies on complex software stacks between 2016 and 2019 (database systems, communication buses and OpenStack). By introducing ENOSLIB, our goal is to gather academic and industrial actors of our community around a library that aggregates everyday experiment-driven research operations. A library that has been already adopted by open-source projects and members of the scientific community thanks to its ease of use and extension (see Section 7 for further details).

Model-Driven Engineering (MDE):

 MDE is a software programming approach that raises the level of abstraction in traditional programming languages by using models with recurring design patterns. MDE simplifies the collaboration between development teams and more broadly promotes compatibility between systems. MDE is increasingly used to help in the development of distributed systems, microservices architectures, and IoT systems but is also leveraged, for instance, in lowcode platforms such as Node-RED to write IoT applications. In 17 two important drawbacks of MDE have been studied. First, as most of the high-abstraction level programming models, when dealing with very big models most MDE frameworks become poorly efficient and slow, thus loosing their initial advantage of speeding up the development process. Second, because of the underlying complexity of MDE approaches the level of confidence in MDE frameworks is usually low. In this paper, we extend the formal transformation engine (transformation being an important part of MDE approaches) CoqTL with a more scalable implementation that is proven equivalent to the initial one, and that is automatically dumped to a Apache Spark code. We have developed a prototype implementation of the refined specification on top of Spark, and have evaluated its performance on a simple case study to assess the speedup our solution can reach.

Software deployment/reconfiguration:

 If speeding and simplifying the development process of distributed systems is of prior importance because of the increasing complexity of geographically distributed infrastructures, speeding and simplifying the deployment of these systems, and enabling and speeding their dynamic evolution through time is another key to making these infrastructures usable. Indeed, because of the scale of both geo-distributed infrastructures and distributed systems (e.g., microservices architectures) manually handling deployments and reconfiguration procedures is an error-prone tedious and complex task, probably impossible in practice. Hence, these procedures need to be automated, or better autonomous. Reconfiguration of systems is a difficult and sensitive procedure during which the system is in between two states, the previous state and the desired state. For this reason, we are interested in two metrics in particular for a few years: the efficiency of the reconfiguration to reach as quick as possible the desired new state; and the safety of the reconfiguration to obtain some formal guarantees on the procedure. In the journal paper 1 we extensively present the reconfiguration formal model Concerto which poses an important pillar in our journey towards autonomous, efficient and safe reconfiguration. Concerto is used to manage the lifecycle of software components and coordinate their reconfiguration operations. Concerto promotes efficiency with a fine-grained representation of dependencies and parallel execution of reconfiguration actions, both within components and between them. In this paper, the elements of the model are described as well as their formal semantics. In addition, we outline a performance model that can be used to estimate the time required by reconfigurations, and we describe an implementation of the model. The evaluation demonstrates the accuracy of the performance estimations, and illustrates the performance gains provided by the execution model of Concerto compared to state-of-the-art systems.

Reifying geo-distribution at the software level:

 Following the results obtained during David Espinel's PhD (see Section 8.1 22, we have investigated how a software stack that has been designed to run on one DC, could benefit from geo-distribution (locality) while dealing with the inherent constraints of wide-area network links without requiring intrusive changes? The admitted approach for several years consists in modifying cloud applications by entangling geo-distribution aspects in the business logic using distributed data stores. However, this makes the code intricate and contradicts the software engineering principle of externalizing concerns. In 15, and then 20, we propose a different approach that relies on the modularity property of microservice applications: (i) one instance of an application is deployed at each edge location, making the system more robust to network partitions (local requests can still be satisfied), and (ii) collaboration between instances can be programmed outside of the application in a generic manner thanks to a service mesh. We validate the relevance of our proposal on a real use-case: geo-distributing OpenStack, a modular application composed of 13 million of lines of code and more than 150 services. We underline these results are at the frontier of the two main research activities of the STACK team (i.e., resource management and programming support).

Blockchain technologies:

 For the past few years, we have been investigating the energy consumption of blockchains, which is particularly criticized. Unfortunately, energy evaluation is often performed with ad hoc tools and specific experimental environments. We have therefore developed BCTMark, a generic framework for benchmarking blockchain technologies on an emulated network in a reproducible way. Leveraging the aforementioned EnosLib library (cf. Section 7.1), BCTMark has enabled us to study a key aspect of the energy consumption of modern blockchains: smart-contract execution 12.

Smart contracts, scripts at the heart of blockchain-based applications, are meant to be available forever once deployed. However, this property comes at a price: the space required to store new contracts continues to increase. We show that over the course of a year, 70% of deployed contracts are not used while they continue to occupy space on the blockchain. To tackle this issue, we have proposed a new protocol to identify and delete unused contracts. Through simulation, based on historical Ethereum data, we have shown that deleting smart contracts after a 90-day period of inactivity could lead to a 66% reduction in the number of contracts stored over a year 5.

Energy efficiency of Industrial IoT devices:

 In 4, we focus on the trade-off between energy-consumption, reliability and latency of different packet forwarding protocols in the wireless Industrial IoT context and propose a new a "braided" forwarding pattern. The aim is to improve existing low-power wireless communication protocols by supporting the strict Quality of Service (QoS) requirements of the industry. Focusing on the IEEE Std 802.15.4-2015 TSCH link-layer standard and the RPL standard at the IETF, we present On-Demand Selection (ODeSe), a novel multi-path routing algorithm, which improves our previous work, the Common Ancestor (CA) algorithms, by selecting the most suitable upward forwarders. Using the Cooja network simulator running the Contiki OS, we compare ODeSe against single-path RPL and multi-path RPL with different alternative parent selection algorithms. The results demonstrate that ODeSe outperforms single-path RPL in terms of reliability, and multi-path RPL in terms of energy consumption while maintaining a 99.14% packet delivery ratio.

Secure and privacy-preserving distributed genetic analyses:

 The increasing availability of sequenced human genomes is enabling health professionals and genomics researchers to well understand the implication of genetic variants in the development of common diseases, notably by means of genome-wide association studies (GWAS) which are very promising for personalized medicine and diagnostic testing. However, the need to handle genetic data from different sources in order to conduct large studies entails multiple privacy and security issues. Actually, classical methods of anonymization are inapplicable for genetic data that are now known to be identifying per se.

We have proposed a novel framework for privacy-preserving collaborative GWAS performed in the cloud. Indeed, our proposal is the first framework which combines a hybrid cloud deployment with a set of four security mechanisms: digital watermarking, homomorphic encryption, meta-data de-identification and the Intel Software Guard Extensions technology. Using these mechanism we can ensure confidentiality of genetic data as well as their integrity. Furthermore, our approach describes meta-data management which has rarely been considered in state-of-the-art propositions despite their importance to genetic analyses. In addition, the new deployment model we have suggested fits with existing infrastructures which makes its integration straightforward. Experimental results of a prototypical implementation on typical data sizes have demonstrated that our solution protocol is feasible and that the framework is practical for real-world scenarios. 14

Privacy-aware distributed FAM:

  FAMD analyses are an important statistical technique that not only enables the visualization of large data but also helps to select subgroups of relevant information for a given patient. While such analyses are well-known in the medical domain, they have to satisfy new data governance constraints if reference data is distributed, notably in the context of large consortia developing the future generation of analyses in the domain of personalised medicine.

We have motivated the use of distributed implementations for FAMD analyses in the context of the development of a personalised medicine application called KITAPP. We have presented a new distribution method for FAMD and evaluated its implementation in a multi-site setting based on real data. Finally we have studied how individual reference data is used to substantiate decision making, while enforcing a high level of usage control and data privacy for patients. 19

Security properties of container technologies:

 Compared to full virtualization, containerization reduces virtualization overhead and resource usage, offers reduced deployment latency and improves reusability. For these reasons, containerization is massively used in an increasing number of applications. However, because containers share a full kernel with the host, they are more vulnerable to attacks that may compromise the host and the other containers on the system. In 13, we present SNAPPY (Safe Namespaceable And Programmable PolicY), a new framework that allows even unprivileged processes such as containers to safely and dynamically enforce in the kernel fine-grained, stackable and programmable eBPF security policies at runtime. This is done by making working coordinately a new LSM (Linux Security Module) Module, a new security Linux namespace abstraction (policy_NS) and eBPF policies enriched with ’dynamic helpers’. This design especially allows to minimize containers’ attack surface.                

Alterway

Participants: Thomas Ledoux, Hiba Awad.

OVHcloud

Participants: Thomas Ledoux, Pierre Jacquet.

Adjunct professorship in Norway

Participants: Hélène Coullon.

10.2.1 FP7 & H2020 projects

10.2.2 Other european programs/initiatives

10.3.1 Ademe

10.3.2 CNRS

10.3.3 ANR

10.3.4 PIA 4

10.3.5 Etoiles Montantes

10.4.1 Interregional activities

10.4.2 Nantes excellency initiative in Medecine and Informatics (NExT)

11.1.1 Scientific events: organisation

11.1.2 Scientific events: selection

11.1.3 Journal

11.1.4 Scientific expertise

• A. Lebre contributed to the Allistene working group dedicated to the definition of “Plan de relance: stratégie d'accélération Cloud et verdissement du numérique” roadmap, March 2021.
• A. Lebre contributed to the working group dedicated to the writing of “Livre blanc Cloud de confiance”, Think tank Digital new deal, May 2021 (further information here.)
• A. Lebre contributes as the IMT representative to the working group dedicated to the organisation and structuring of the PEPR Cloud action (Budget overall: 56M€).
• A. Lebre contributed to the writing of “Livre blanc sur l'Internet des objets (IoT)”, Inria, Dec 2021.
• T. Ledoux was a reviewer for the ADEME call Perfecto 2021.
• J.-M. Menaud was an expert for the 2021 CIR evaluation

11.1.5 Research administration

• A. Lebre is a member of the executive and architect committees of the Grid’5000 GIS (Groupement d’intérêt scientifique).
• A. Lebre is a co-director of the <I/O> Lab, a joint lab between Inria and Orange Labs.
• A. Lebre is a member of the scientific committee of the joint lab between Inria and Nokia Bell Labs.
• H. Coullon is vice-president of the French ACM SIGOPS group.
• H. Coullon is co-chair of the working group YODA (trustworthY and Optimal Dynamic Adaptation) in the national reasearch group GDR GPL (software engineering and languages).
• H. Coullon is co-chair of the working group CyclOps (DevOps, deployment and reconfiguration, lifecycle etc.) of the joint laboratory between Inria and Orange (<I/O> Lab).
• J. Noyé is the deputy head of the Automation, Production and Computer Sciences Department of IMT Atlantique.
• M. Südholt is a deputy director of the graduate school MathSTIC that covers the two regions Pays de la Loire and Brittany. He is responsible for the management of doctoral studies in the Nantes region.
• M. Südholt is co-chair of the working group Security of the joint laboratory between Inria and Orange (<I/O> Lab).
• J.-M. Menaud is the organizer of "Pôle Science du Logiciel et des Systèmes Distribués" in Laboratoire des Sciences du Numérique à Nantes (LS2N).
• J.-M. Menaud is co-chair of the working group Energy of the joint laboratory between Inria and Orange (<I/O> Lab)
• J.-M. Menaud is in charge of the academic relations for the Automation, Production and Computer Sciences Department of IMT Atlantique.
• J.-M. Menaud is involved in the GIS (Groupement d’Intérêt Scientifique) VITTORIA (VIrTual inTegrative Oncology Research and InnovAtion) and PERLE (Pôle d’Excellence de la Recherche Ligérienne en Energie).

11.2.1 Teaching

• T. Ledoux is the head of the apprenticeship program in Software Engineering FIL. This 3-year program leads to the award of a Master degree in Software Engineering from the IMT Atlantique.
• H. Coullon is responsible for the Computer Science domain of the new apprenticeship program in Industry 4.0 (FIT) of IMT Atlantique. This 3-year program leads to the award of a Master degree in Industry 4.0 from the IMT Atlantique.
• T. Ledoux has been the head of the Filière informatique nantaise since Sept. 2020. This entity, created by the University of Nantes, Centrale Nantes and IMT Atlantique, aims to bring together the main players in Computer Science training in Nantes to ensure a coherent and ambitious training offer that meets the present and future challenges of Computer Science. It is organized around a Council made up of representatives from the academic and socio-economic worlds.
• J. Noyé is the deputy head of the Automation, Production and Computer Sciences Department of IMT Atlantique.
• M. Südholt is the representative for MSc-level and PhD-level studies of the API department of IMT Atlantique.

11.2.2 Supervision

• PhD: Marie Delavergne, director: A. Lebre.
• PhD: David Espinel, director: A. Lebre (defended Sept. 2021).
• PhD: Geo Johns Anthony, director: A. Lebre (since Apr. 2021).
• PhD: Dimitri Saingre, advisor: T. Ledoux, director: J-M. Menaud (defended Dec. 2021).
• PhD: Jolan Philippe, advisors: H. Coullon, M. Tisi (NaoMod), director: G. Sunye (NaoMod).
• PhD: Maxime Belair, advisor: S. Laniepce (Orange Labs), director: J.-M. Menaud (defended Dec. 2021).
• PhD: Wilmer Garzon, advisor: M. Südholt.
• PhD: Sirine Sayadi, advisor: M. Südholt.
• PhD: Pierre Jacquet, advisor: T. Ledoux (since Oct. 2021).
• PhD: Hiba Awad, advisor: T. Ledoux (since Nov. 2021).
• PhD: Antoine Omond, advisor: H. Coullon, director: T. Ledoux (since Dec. 2021).
• Postdoc: Rémy Pottier, advisor: J.-M. Menaud (until Oct. 2021).
• Postdoc: Simon Robillard, advisor: H. Coullon (until June 2021).
• Engineer: Emile Cadorel, advisor: H. Coullon (until Jan. 2021).
• Postdoc: Brice Nédelec, advisor: A. Lebre (from Sept. to Dec. 2021).
• Post-doc: Abdelghani Alidra, advisor: T. Ledoux (since May 2021).
• Engineer: Ronan-Alexandre Cherrueau, advisor: A. Lebre (until July 2021).
• Engineer: Brice Nédelec, advisor: T. Ledoux (until Mar. 2021).

11.2.3 Juries

• A. Lebre was a reviewer of the Phd Committee of Alessio Diamanti, “A novel network automation architecture: from anomaly detection to dynamic reconfiguration”, Conservatoire National des Arts et Metier Paris, Dec 2021.
• T. Ledoux was a member of the PhD committee of David Espinel, “Distributing connectivity management in Cloud-Edge infrastructures using SDN-based approaches”, IMT Atlantique, Sep. 07, 2021.
• T. Ledoux was a reviewer of the PhD committee of Zakaria Ournani, “Software Eco-Design: Investigating and Reducing the Energy Consumption of Software”, Univ. Lille, Nov. 08, 2021.
• H. Coullon was a member of the PhD committee of Tanissia DJEMAI, "Placement optimisé de services dans les architectures Fog Computing et Internet of Things sous contraintes d’énergie, de QoS et de mobilité" (French), Université de Toulouse, Feb. 03, 2021.
• H. Coullon was a member of two selection committees for associate professors at the university of Rennes and the university of Bordeaux.

11.3.1 Articles and contents

• T. Ledoux, “La tête dans les nuages”, Atlanstic (online), Jan. 2021.
• A. Lebre contributed to the article “La double revolution du Edge Computing”, Les echos, Sept. 2021.
• T. Ledoux, `IMT Atlantique et Alter Way veulent inventer le circuit court de la donnée”, POC Media, Sept. 2021.

11.3.2 Education

T. Ledoux is co-leader of the Ecolog initiative. This national project is the result of the association between the Institut du Numérique Responsable and the IT sector in Nantes, which includes the University of Nantes, IMT Atlantique and Centrale Nantes. Its ambition is to create a body of training courses available in open source and dedicated to green computing.

Lately, R.-A. Koustiamanis has been taking part in the development of two MOOCs related to the Industrial IoT, one specific to IMT Atlantique and the other one as a partnership of the German-French Academy for the Industry of the Future.

International journals

• 6 articleA.Amaury Bruniaux, R.-A.Remous-Aris Koutsiamanis, G.Georgios Papadopoulos and N.Nicolas Montavont. Defragmenting the 6LoWPAN Fragmentation Landscape: A Performance Evaluation.Sensors215March 2021, 1-20
  HALDOIback to text
• 7 articleM.Maverick Chardet, H.Hélène Coullon and S.Simon Robillard. Toward Safe and Efficient Reconfiguration with Concerto.Science of Computer Programming203March 2021, 1-31
  HALDOI
• 8 articleR.-A.Ronan-Alexandre Cherrueau, M.Marie Delavergne, A.Alexandre Van Kempen, A.Adrien Lebre, D.Dimitri Pertin, J.Javier Rojas Balderrama, A.Anthony Simonet and M.Matthieu Simonin. EnosLib: A Library for Experiment-Driven Research in Distributed Computing.IEEE Transactions on Parallel and Distributed Systems336June 2022, 1464-1477
  HALDOI
• 9 articleD.David Espinel Sarmiento, A.Adrien Lebre, L.Lucas Nussbaum and A.Abdelhadi Chari. Decentralized SDN Control Plane for a Distributed Cloud-Edge Infrastructure: A Survey.Communications Surveys and Tutorials, IEEE Communications Society231February 2021, 256-281
  HALDOI
• 10 articleD.David Hauweele, R.-A.Remous-Aris Koutsiamanis, B.Bruno Quoitin and G.Georgios Papadopoulos. Thorough Performance Evaluation & Analysis of the 6TiSCH Minimal Scheduling Function (MSF).Journal of Signal Processing SystemsJune 2021
  HALDOIback to text
• 11 articleT.Tomas Lagos Jenschke, R.-A.Remous-Aris Koutsiamanis, G.Georgios Papadopoulos and N.Nicolas Montavont. ODeSe: On-Demand Selection for multi-path RPL networks.Ad Hoc Networks1142021, 102431
  HALDOI
• 12 articleD.Dimitri Saingre, T.Thomas Ledoux and J.-M.Jean-Marc Menaud. Measuring performances and footprint of blockchains with BCTMark: a case study on Ethereum smart contracts energy consumption.Cluster Computing2021
  HALDOIback to text

International peer-reviewed conferences

• 13 inproceedingsM.Maxime Belair, S.Sylvie Laniepce and J.-M.Jean-Marc Menaud. SNAPPY: Programmable Kernel-Level Policies for Containers.SAC 2021: 36th ACM/SIGAPP Symposium On Applied ComputingGwangju / Virtual, South KoreaMarch 2021, 1636–1645
  HALDOIback to text
• 14 inproceedingsF.-z.Fatima-zahra Boujdad, D.David Niyitegeka, R.Reda Bellafqira, G.Gouenou Coatrieux, E.Emmanuelle Génin and M.Mario Südholt. A Hybrid Cloud Deployment Architecture For Privacy-preserving Collaborative Genome-Wide Association Studies.ICDF2C - 12th EAI International Conference on Digital Forensics & Cyber CrimeSingapour, SingaporeDecember 2021
  HALback to textback to text
• 15 inproceedingsR.-A.Ronan-Alexandre Cherrueau, M.Marie Delavergne and A.Adrien Lebre. Geo-Distribute Cloud Applications at the Edge.EURO-PAR 2021 - 27th International European Conference on Parallel and Distributed ComputingLisbon, PortugalAugust 2021, 1-14
  HALDOIback to text
• 16 inproceedingsK.Karim Manaouil and A.Adrien Lebre. Kubernetes WANWide: a Deployment Scenario to Expose and Use Edge Computing Resources?PDP 2021 - 29th Euromicro International Conference on Parallel, Distributed and Network-Based ProcessingValladolid, SpainIEEEMarch 2021, 193-197
  HALDOIback to text
• 17 inproceedingsJ.Jolan Philippe, M.Massimo Tisi, H.Hélène Coullon and G.Gerson Sunyé. Executing Certified Model Transformations on Apache Spark.SLE 2021: 14th ACM SIGPLAN International Conference on Software Language EngineeringSLE 2021: Proceedings of the 14th ACM SIGPLAN International Conference on Software Language EngineeringChicago IL, United StatesNovember 2021, 36-48
  HALDOIback to text
• 18 inproceedingsD.Dimitri Saingre, T.Thomas Ledoux and J.-M.Jean-Marc Menaud. The cost of immortality: A Time To Live for smart contracts.ISCC 2021 - 26th IEEE Symposium on Computers and CommunicationsAthènes, GreeceIEEESeptember 2021
  HAL
• 19 inproceedingsS.Sirine Sayadi, E.Estelle Geffard, M.Mario Südholt, N.Nicolas Vince and P.-A.Pierre-Antoine Gourraud. Secure distribution of Factor Analysis of Mixed Data (FAMD) and its application to personalized medicine of transplanted patients.AINA 2021: 35th International Conference on Advanced Information Networking and ApplicationsToronto, CanadaApril 2021
  HALDOIback to textback to text

Conferences without proceedings

• 20 inproceedingsM.Marie Delavergne, R.-A.Ronan-Alexandre Cherrueau and A.Adrien Lebre. A service mesh for collaboration between geo-distributed services: the replication case.AMP 2021 - Workshop on Agility with Microservices ProgrammingOnline, FranceJune 2021, 1-8
  HALback to text

Doctoral dissertations and habilitation theses

• 21 thesisM.Maxime Bélair. Defense against attacks with regard to new virtualisation kinds.Ecole nationale supérieure Mines-Télécom AtlantiqueDecember 2021
  HAL
• 22 thesisD.David Espinel Sarmiento. Distributing connectivity management in Cloud-Edge infrastructures using SDN-based approaches.Ecole nationale supérieure Mines-Télécom AtlantiqueSeptember 2021
  HALback to textback to text
• 23 thesisD.Dimitri Saingre. Understanding the energy consumption of blockchain technologies : a focus on smart contracts.Ecole nationale supérieure Mines-Télécom AtlantiqueDecember 2021
  HAL

Reports & preprints

• 24 reportA.Adrien Lebre, B.Brice Nédelec and A.Alexandre Van Kempen. AS-cast: Lock Down the Traffic of Decentralized Content Indexing at the Edge.RR-9418Inria RennesSeptember 2021, 1-21
  HALback to text