Title: An autonomic management system for IoT platforms based on data analysis tasks

Authors: Clovis Anicet Ouedraogo; Jose Aguilar; Christophe Chassot; Samir Medjiah; Khalil Drira

Addresses: CNRS-LAAS, 31400 Toulouse, France ' CEMISID, Universidad de Los Andes, Mérida, Venezuela; GIDITIC, Universidad EAFIT, Medellin, Colombia ' CNRS-LAAS, 31400 Toulouse, France; INSA, F-31400 Toulouse, France ' CNRS-LAAS, 31400 Toulouse, France; UPS, Université de Toulouse, Toulouse, France ' CNRS-LAAS, 31400 Toulouse, France

Abstract: In this work, we propose an autonomic management system (AMS) for the internet of things (IoT) platforms, which uses the concept of autonomic cycle of data analysis tasks to improve and maintain the performance in the IoT platforms. The concept of 'autonomic cycle of data analysis tasks' is a type of autonomous intelligent supervision that allows reaching strategic objectives around a given problem. In this paper, we propose the conceptualisation of the architecture of an AMS composed by an autonomic cycle to optimise the quality of services (QoS), and to improve the quality of experiences (QoE), in IoT platforms. The autonomous cycle detects and discovers the current operational state in the IoT platform and determines the set of tasks to guarantee a given performance (QoS/QoE). This paper presents the details of the architecture of the AMS (components, knowledge models, etc.), and its utilisation in two case studies: in a typical application in an IoT context, and in a tactile internet system.

Keywords: autonomic computing; internet of things; IoT; data analysis tasks; autonomic management system; AMS; quality of experiences; QoE.

DOI: 10.1504/IJCNDS.2022.125362

International Journal of Communication Networks and Distributed Systems, 2022 Vol.28 No.5, pp.575 - 599

Received: 29 May 2021
Accepted: 21 Sep 2021
Published online: 07 Sep 2022 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article