Projects
Research and projects development
DRUID-NET: Edge Computing Resource Allocation for Dynamic Networks (CHIST-ERA 2018) (May 2020 – April 2023)
Investigators: Nikolaos Athanasopoulos (Queen’s University Belfast), Raphael Jungers (UCLouvain), Aris Leivadeas (ÉTS), Nathalie Mitton (INRIA Lille), Symeon Papavassiliou (National Technical University of Athens)
eDge computing ResoUrce allocatIon for Dynamic NETworks (DRUID-NET) is a three-year project that falls under the CHIST-ERA program. The objective of DRUID-NET is to design and develop a new generation of workload estimators – performance models that efficiently capture the time-varying nature of IoT-based applications, dynamic networks and edge computing infrastructure. Building on this framework, DRUID-NET aspires to develop novel resource-allocation mechanisms that explicitly take into account service differentiation and context-awareness, and most importantly, provide formal guarantees for well-defined QoS/QoE metrics. Our research also goes well beyond the state-of-the-art in the design of control algorithms for cyber-physical systems (CPS) by incorporating resource allocation mechanisms into the decision strategy itself. The DRUID-NET project proposes a new generation of controllers, driven by a co-design philosophy, both in the network and in computing resources utilization. This paradigm has the potential to cause a quantum leap in crucial fields in engineering (e.g.: Industry 4.0, collaborative robotics, logistics, multi-agent systems, etc.). To achieve these breakthroughs, we utilize and combine tools from fields where the consortium possesses leading expertise on an international scale, namely Automata and Graph theory, Machine Learning, Modern Control Theory and Network Theory.
RAMOSE: Resource provisioning Automation, Management and Orchestration for Service chain Embedding in cloud environments
Investigator: Nadjia Kara
RAMOSE is a four-year project intended to design, implement and test a framework for automating the deployment and the management of service function chains (SFCs) while meeting application performance requirements. This framework enables efficient configuration, management and control of SFCs using various theories and AI techniques (e.g.: AI techniques, graph theory, game theory). It provides mechanisms and tools for: 1) dynamic resource deployment and adaptation; 2) service configuration and management; 3) resource monitoring and prediction.
Principal investigator: Kaiwen Zhang
Blockchain-based distributed ledger technology (DLT) processes and disseminates transaction events in a decentralized network. Our objective is to achieve the right balance of decentralization, consistency and scalability to meet the requirements of important domains, such as eHealth, IoT and Smart Grids.
5G/B5G and IoT
Investigator: Diala Naboulsi
Future mobile networks, including 5G and 6G, are envisioned to be multiservice networks that provide support for various vertical markets. Corresponding use cases present very heterogeneous and diverse requirements. Existing mobile network technologies lack the flexibility that allows them to accommodate these use cases efficiently. As a result, network slicing has recently emerged as a key concept that allows this challenge to be tackled. It does so by enabling the creation of network slices, consisting of logical networks, on top of a single physical network, on a per-service basis, while considering their respective requirements. Network slicing has received significant attention in the literature over the past few years. However, realizing the vision of a slicing-based network remains a difficult goal to achieve today.
In this proposal, we aim to realize the vision of a dynamic slicing-based mobile network that can create per-service network slices on the fly while satisfying the requirements of each service. Considering this target, our goal is to answer the following research questions: i) How can we jointly optimize the allocation of core network and radio access network resources for network slicing? ii) How can we manage end /device mobility in network slices? iii) How can we realistically model traffic dynamics for the evaluation of network slicing solutions? iv) What components are included in an efficient end-to-end network slicing architecture?
Principal investigators: Ghizlane El Boussaidi and Sègla Kpodjedo
This project aims to integrate an IoT solution into the legacy system of an industrial partner. Modernizing such a large software system is a demanding and costly task that requires knowledge and understanding of the system architecture. The specific goals of this project are two-fold: 1) Establish scenarios and quality attributes that stem from the integration of IoT devices and that must be supported by the system’s architecture, and 2) Assess the software architecture of this system based on these scenarios and quality attributes. The outcome of this case study will help to synthesize a more systematic approach to architectural analysis within the context of an IoT-based modernization process.
Principal investigator: Ghizlane El Boussaidi
Despite the progress made by existing approaches in terms of managing and documenting design decisions, we still lack effective support for enforcing architectural decisions (e.g.: architectural styles, tactics and patterns) during the development and evolution of systems. In fact, recent studies reveal that developers do not always understand the impact of their changes on architecture, especially within the context of distributed and heterogeneous systems, such as IoT applications. In light of this, the goal of this research is to study common existing IoT development frameworks and technologies in order to determine how they support common design decisions, and to use this knowledge to develop techniques and tools that enforce these decisions and prevent architectural erosion.
Investigator: Nadjia Kara
The objective of this project is to develop new mechanisms to support the efficient deployment of a fleet of UAVs equipped with rotating cameras over a large-scale geographical area for monitoring and surveillance purposes. These mechanisms include: 1) methods for stabilization, control and guidance of UAVs in altitude and attitude, using AI and Edge computing techniques; 2) mission planning and path tracking; 3) UAV swarm formation; 4) customizable secure communication system for UAVs.
Distributed communication systems
Principal investigator: Kaiwen Zhang
Modern applications (e.g.: social networks, sensor networks) require advanced analytics over real-time streams of events. We strive to support these expressive features without compromising performance or scalability by integrating lightweight distributed processing techniques into the publish/subscribe communication substrate.
Principal investigator: Kaiwen Zhang
Networked games require low-latency, reliable and cheat-resistant communication between players. Our goal is to develop scalable infrastructures for highly-interactive online games. By leveraging game semantics, such as player profiles and the virtual environment, we can achieve the right trade-off between consistency and performance.
Support for system design
Prinicpal investigator: Ghizlane El Boussaidi
Co-investigators: Abdelouahed Gherbi and Sègla Kpodjedo
In this project, we are seeking to leverage model-driven techniques for building an integrated approach to support the avionics software industry in specifying and collecting requirements and test evidence needed for safety-critical software in general, and for an RTOS (real-time operating system) product line in particular. The expected results are methods and tools that support model-based development, specification of requirements, test automation, product line development and configuration management. These methods and tools will improve the industrial partner’s productivity and strengthen its position in the development of avionics software.
Testbeds
This demo shows a model for the calculation of Bitcoin performance using block size and a default number of connections per node. The analytical model has been validated through simulation and compared to real data measured from the Bitcoin network. Although the throughput of Bitcoin can be increased by choosing a larger size for blocks, this can cause a significant increase in block propagation time. The delay can be reduced by increasing the average default number of connections per node, but the drawback is increased traffic overhead in the network.