KHIZAR Sadia

PhD graduated
Team : NPA
    Sorbonne Université - LIP6
    Boîte courrier 169
    Couloir 26-00, Étage 1, Bureau 120
    4 place Jussieu
    75252 PARIS CEDEX 05
    FRANCE

Tel: +33 1 44 27 88 39, Sadia.Khizar (at) nulllip6.fr
https://lip6.fr/Sadia.Khizar

Supervision : Marcelo DIAS DE AMORIM

Co-supervision : CONAN Vania

Leveraging mobile devices beyond the edge toward task offloading

The pervasiveness of mobile devices equipped with internet connectivity and positioning systems, ranging from vehicles, mobile phones to smartwatches and fitness bracelets, leads us to regard them as a valuable resource to leverage. In this thesis, we tackle the use of mobile devices from a new perspective. We consider the extension of the capacity of the MEC by using the available resources of mobile devices beyond the edge of the infrastructure network. The goal is to leverage their untapped resources to process computation on behalf of the MEC in a distributed way. It is fundamental for the MEC orchestrator to be aware of its operating environment to rely on mobile nodes. Therefore, MEC needs to estimate if they have enough computing, storage, and battery resources, but also for how long. In the first part of the thesis, we have focused on the temporal availability of beyond-the-edge resources. We chose to investigate the co-location of terminals and analyze their persistence in a cell. Then, we turn our attention to task allocation. We shift the focus on the spatio-temporal aspect by quantifying the resources that a cell can provide to perform a MEC task. We estimate the potential amount of computational tasks performed by nodes based on the cumulative presence time in a given cell and a given completion delay. Results provide insight into the possibilities of offloading computing tasks on mobile nodes. Furthermore, it allows knowing the locations where it is advisable to offload tasks and the time duration of tasks offloadable. In addition, the MEC may have a specific guarantee of having a sufficient number of mobile devices to perform a task. Finally, the results give an overview of additional tasks feasible based on the delay completion tolerance that the MEC is willing to accept.

Defence : 02/15/2022 - 14h30 - Campus Pierre et Marie Curie, salle Jacques Pitrat (25-26/105)

Jury members :

Nadjib Ait Saadi, Prof. Université Versailles Saint-Quentin [Rapporteur]
André-Luc Beylot, Prof. ENSEEIHT [Rapporteur]
Anne Fladenmuller, Prof. Sorbonne Université
Thi-Mai-Trang Nguyen, Assoc. Prof. Sorbonne Université
Vania Conan, DR Thales
Marcelo Dias de Amorim, DR CNRS

2019-2022 Publications