EN FR
Homepage Inria website


Section: Partnerships and Cooperations

European Initiatives

FP7 & H2020 Projects

LightKone
  • Title: Lightweight Computation for Networks at the Edge

  • Programm: H2020-ICT-2016-2017

  • Duration: January 2017 - December 2019

  • Coordinator: Université Catholique de Louvain

  • Partners:

    • Université Catholique de Louvain (Belgium)

    • Technische Universitaet Kaiserslautern (Germany)

    • INESC TEC - Instituto de Engenharia de Sistemas e Computadores, Tecnologia e Ciencia (Portugal)

    • Faculdade de Ciencias E Tecnologiada Universidade Nova de Lisboa (Portugal)

    • Universitat Politecnica De Catalunya (Spain)

    • Scality (France)

    • Gluk Advice B.V. (Netherlands)

  • Inria contact: Marc Shapiro

  • The goal of LightKone is to develop a scientifically sound and industrially validated model for doing general-purpose computation on edge networks. An edge network consists of a large set of heterogeneous, loosely coupled computing nodes situated at the logical extreme of a network. Common examples are networks of Internet of Things, mobile devices, personal computers, and points of presence including Mobile Edge Computing. Internet applications are increasingly running on edge networks, to reduce latency, increase scalability, resilience, and security, and permit local decision making. However, today’s state of the art, the gossip and peer-to-peer models, give no solution for defining general-purpose computations on edge networks, i.e., computation with shared mutable state. LightKone will solve this problem by combining two recent advances in distributed computing, namely synchronisation-free programming and hybrid gossip algorithms, both of which are successfully used separately in industry. Together, they are a natural combination for edge computing. We will cover edge networks both with and without data center nodes, and applications focused on collaboration, computation, and both. Project results will be new programming models and algorithms that advance scientific understanding, implemented in new industrial applications and a startup company, and evaluated in large-scale realistic settings.