Abstract:
Fog computing is an emerging technology that extends the cloud and brings
the resources closer to the end devices to achieve better performance in
latency-sensitive application. Fog computing still lacks the standardization
in terms of simulation environment and architecture. Several fog simulators
has been developed and proposed previously. Most of the existing fog
simulators ignore core network characteristics like error rate, packet loss,
bandwidth etc. There are a number of fog simulators available today, among
which a few are open-source, whereas rest are commercially available. The
existing fog simulators mainly focus on a number of devices that can be simulated.
Generally, the existing simulators are more inclined toward sensors'
con gurations, where sensors generate raw data and fog nodes are used to intelligently
process the data before sending to back-end cloud or other nodes.
Therefore, these simulators lack network properties and assume reliable and
error-free delivery on every service request. Moreover, no simulator allows researchers
to incorporate their own fog nodes management algorithms, such as
scheduling. In existing work, device handover is also not supported. In this
paper, we propose a new fog simulator called FogNetSim++1 that provides
users with detailed con guration options to simulate a large fog network.
It enables researchers to incorporate customized mobility models, fog node
scheduling algorithms, and manage handover mechanisms. In our evaluation
setup, a tra c management system is evaluated to demonstrate the scalability
and e ectiveness of proposed simulator in terms of CPU, and memory
usage. We have also benchmarked the network parameters such as execution
delay, packet error rate, handovers, and latency.
