Scheduling And Qos Management On Fog Computing Architecture

Universities and Institutes of France
September 29, 2022
Offerd Salary:Negotiation
Working address:N/A
Contract Type:Temporary
Working Time:Full time
Working type:N/A
Job Ref.:N/A
  • Organisation/Company: CNRS
  • Research Field: Computer science Mathematics › Algorithms
  • Researcher Profile: First Stage Researcher (R1)
  • Application Deadline: 29/09/2022 23:59 - Europe/Brussels
  • Location: France › TOULOUSE
  • Type Of Contract: Temporary
  • Job Status: Full-time
  • Hours Per Week: 35
  • Offer Starting Date: 01/10/2022
  • LAAS CNRS - SARA Team (Services et Architectures pour Réseaux Avancés)


    The explosion of the volume of data exchanged within today's IT systems, due to an increasingly widespread use and by an ever wider audience (large organizations, companies, general public etc.), has led to questioning the architectures used until now. Indeed, for the past few years, Fog computing 1, which extends the Cloud computing paradigm to the edge of the network, has been developing steadily, offering more and more possibilities and thus expanding the field of Internet of Things applications. The management of these new architectures, involving a large number of devices and heterogeneous applications, and potentially large volumes of data to be processed, is a challenge to propose innovative and efficient solutions to improve the availability, stability, fluidity, security and efficiency of the services and applications deployed.

    Problems addressed :

    The set of problems of the thesis will cover several aspects: Task scheduling in a FOG environment, Energy management at different levels (potentially storage-limited fog equipment, intermediate compute node, energy-intensive cloud compute node), Quality of Service management of FOG services (QoS FOG), Understanding and development of one or more learning and prediction methods to help the decision making of a meta-scheduler, finally, experiments will have to be carried out on a large scale distributed platform such as G5K (Grid'5000). The simulation aspect will also be important in order to propose a new module for BatSim (FOG extension of the French simulator "simgrid") integrating the meta-scheduler as well as the learning and prediction methods developed.

    The field of study being very vast and gathering various aspects of research, an effort of modeling of the studied system will have, above all, to be proposed after a detailed study of the works of the state of the art in the field of the Fog Computing: QoS & Energy, centralized/decentralized Scheduling, method of AI directed scheduling, heuristics.

    One of the problems will be to study and propose a task scheduling solution, under energy and Quality of Service (QoS) constraints, integrating a learning phase on the intrinsic characteristics of the tasks arriving at the services. Thus, the choice of the algorithm or heuristic to schedule each task or set of tasks will be guided by a prediction resulting from the chosen machine learning method (or deep learning). We will thus speak of a centralized "meta-scheduler".

    In the context of this thesis, the problems of QoS management and energy consumption will have to be studied taking into account the intrinsic characteristics of the domain. For the QoS, a part of study and definition of relevant and measurable non-functional parameters will have to be done. On the energy side, the work will have to focus on the aspect of managing the quantities of energy available in the whole platform. With heterogeneous equipments, for example, either limited in computing capacity or limited in energy storage, scheduling compromises will have to be found in order to satisfy "as well as possible" the constraints and objectives of QoS.

    Another issue will be to propose a module for predicting the behavior of the platform: behavior in terms of usage over time (workload), state and quantity of energy of equipment depending on storage, availability and rate of use of services over the time horizon considered or, for example, a need to move certain services in order to satisfy certain QoS criteria. Scaling analysis studies will have to be done in order to evaluate the efficiency of this module.

    Given the low number of real platforms at the moment, it will also be essential to focus on the elaboration of a use case scenario in order to evaluate the developed methods in conditions that are close to the real world. Also, the experiments envisaged could be carried out on the distributed platform G5K (GRID'5000).

    Translated with www. (free version)

    Web site for additional job details

    https: //

    Required Research Experiences
  • Computer science

  • None

  • Mathematics › Algorithms

  • None

    Offer Requirements
  • Computer science: Master Degree or equivalent

    Mathematics: Master Degree or equivalent

  • FRENCH: Basic

    Contact Information
  • Organisation/Company: CNRS
  • Department: Laboratoire d'analyse et d'architecture des systèmes
  • Organisation Type: Public Research Institution
  • Website: https:// www.
  • Country: France
  • City: TOULOUSE
  • From this employer

    Recent blogs

    Recent news