Details

Autor(en) / Beteiligte

• Aburukba, Raafat O.
• AliKarrar, Mazin
• Landolsi, Taha
• El-Fakih, Khaled

Titel

Scheduling Internet of Things requests to minimize latency in hybrid Fog–Cloud​ computing

Ist Teil von

• Future generation computer systems, 2020-10, Vol.111, p.539-551

Ort / Verlag

Elsevier B.V

Erscheinungsjahr

2020

Quelle

Access via ScienceDirect (Elsevier)

Beschreibungen/Notizen

• Delivering services for Internet of Things (IoT) applications that demand real-time and predictable latency is a challenge. Several IoT applications require stringent latency requirements due to the interaction between the IoT devices and the physical environment through sensing and actuation. The limited capabilities of IoT devices require applications to be integrated in Cloud and Fog computing paradigms. Fog computing significantly improves on the service latency as it brings resources closer to the edge. The characteristics of both Fog and Cloud computing will enable the integration and interoperation of a large number of IoT devices and services in different domains. This work models the scheduling of IoT service requests as an optimization problem using integer programming in order to minimize the overall service request latency. The scheduling problem by nature is NP-hard, and hence, exact optimization solutions are inadequate for large size problems. This work introduces a customized implementation of the genetic algorithm (GA) as a heuristic approach to schedule the IoT requests to achieve the objective of minimizing the overall latency. The GA is tested in a simulation environment that considers the dynamic nature of the environment. The performance of the GA is evaluated and compared to the performance of waited-fair queuing (WFQ), priority-strict queuing (PSQ), and round robin (RR) techniques. The results show that the overall latency for the proposed approach is 21.9% to 46.6% better than the other algorithms. The proposed approach also showed significant improvement in meeting the requests deadlines by up to 31%.