A PROPOSED SMART RESOURCE SCHEDULING SYSTEM IN FOG COMPUTING

One of the most remarkable innovative ideas in recent technology
advancement is fog computing. It addresses a number of cloud computing
shortcomings by bringing computation, storage, and actual services closer to end
users. However, the majority of fog devices are resource restricted. As a result,
without efficient resource scheduling, leveraging the benefits of fog computing
is challenging.
The idea of resource scheduling is to select the most suitable resources
for the applicant in order to accomplish the best scheduling target. The majority
of the recent studies have been on expanding the number of fog nodes in order to
improve the available resources. This has led to the emergence of a number of
other problems such as increasing the cost and amount of energy consumed.
Therefore, the aim of this thesis is to propose a scheduling system that will
schedule the available resources in the fog layer and distribute them according to
the required tasks without the need to increase these nodes. It has an extra layer,
Master Fog (MF), between the cloud and general-purpose fogs, referred to as
Citizen Fog (CF). The MF is responsible of task execution in CFs and the cloud.
The Comparative Attributes Algorithm (CAA) is used to prioritize jobs, and the
Linear Attribute Summarized Algorithm (LASA) is used to choose the most
available CF with the highest computational resources. To analyze the proposed
solution, iFogSim was used to create a simulation architecture and environment.
The final results indicate a noticeable scheduling of available sources
represented by an increase in bandwidth by 14%, in addition to an increase in
processing speed by 34%. On the other hand, there was a reduction in RAM
consumed by approximately 14%, in addition to a reduction in energy consumed
by approximately 14%.