![](bg2.png)
2 RELATED WORK
We can find several survey papers in the field of Cloud
Computing and Fog Computing of tools supporting
modelling and simulation. Concerning the proper-
ties and modelling of Fog Computing, Puliafito et al.
(Puliafito et al., 2019) presented a survey highlight-
ing and categorizing the properties of Fog Comput-
ing, and investigated the benefits of applying fogs to
support the needs of IoT applications. They intro-
duced six IoT application groups exploiting fog capa-
bilities, and gathered fog hardware and software plat-
forms supporting the needs of these IoT applications.
Markus et al. (Markus and Kertesz, 2019) focused
on available cloud, IoT and fog simulators, and com-
pared them according to several metrics such as soft-
ware metrics and general characteristics. Concerning
fog simulation, they introduced and classified 18 sim-
ulators. We selected five recent fog simulators, and
briefly compared them in Table 1. We noted their
base simulator, publication date and type for their cat-
egorization. The network type simulators usually fo-
cus on low-level network interaction between entities
such as routers, switches and nodes, but less suitable
for the higher level of abstraction (e.g. virtual ma-
chines), whilst event-driven type simulators are more
general and usually lack implemented the network op-
erations or only support minimal network traffic sim-
ulation, but they are easier to be used for accurate rep-
resentation of higher level system components. We
also summarized the number of literature search re-
sults (i.e. hits) performed in Google Scholar
1
, and we
summed the number of citations of the top five rele-
vant hits.
DISSECT-CF-Fog is based on DISSECT-CF, and
a direct extension of the DISSECT-CF-IoT simulator
(Markus et al., 2017), also developed by the authors.
The base simulator is able to model cloud environ-
ments and supports energy measurements of physical
resources. The extended version supports the mod-
elling of IoT systems and its communications. The
whole software is fully configurable, and follows a
hierarchical structure.
EdgeCloudSim (Sonmez et al., 2017) is a
CloudSim extension with the main capabilities of net-
work modelling, including extensions for WLAN,
WAN and device mobility. The developers of this tool
aimed to respond to the disadvantage of the simple
network model of iFogSim by introducing network
load management and content mobility to this sim-
ulator.
The FogNetSim++ (Qayyum et al., 2018) is built
1
Google Scholar is available at: https://scholar.google.com
Accessed in September, 2019.
on the OMNeT++ discrete event simulator, which fo-
cuses on network simulation. This extension pro-
vides configuration options for fog network manage-
ment including node scheduling and selection. It is
also able to model different communication protocols,
such as MQTT or CoAP, and different mobility mod-
els.
One of the most applied and referred fog simula-
tors is iFogSim (Gupta et al., 2016), which is based
on CloudSim. iFogSim can be used to simulate cloud
and fog systems using the sensing, processing and ac-
tuating model. It is able to model cloud and fog de-
vices with certain resource parameters. Sensors and
actuators can also be managed represented by a Tu-
ple. There are dedicated modules for processing and
data-flows.
DockerSim (Nikdel et al., 2017) aims to support
the analysis of container-based SaaS systems in sim-
ulated environments. It is based on the iCanCloud
network simulator, this extension can model container
behaviour, network, protocol and OS process schedul-
ing behaviour.
Though all of these simulators would be interest-
ing to be further analysed, after performing a quick
pre-evaluation we found that iFogSim and DISSECT-
CF-Fog are the most mature and documented solu-
tions, and we also took into account a literature search
result and number of citations for our decision. We
also considered numerous iFogSim extensions, which
have appeared in the last few years, and the sup-
port for novel functions or properties of Fog Comput-
ing (as proposed by a recent survey in (Markus and
Kertesz, 2019)). Unfortunately, only a few of those
extensions were published with available source code,
thus our goal was to make a comparison with the orig-
inal version of iFogSim. A former simulator compar-
ison by Mann (Mann, 2018) also had an effect on our
decision, which addressed the core of these simulators
(namely CloudSim and DISSECT-CF).
3 FOG MODELLING IN iFogSim
AND DISSECT-CF-Fog
The CloudSim-based extensions (e.g. iFogSim
or EdgeCloudSim) are often used for investigating
Cloud and Fog Computing approaches, and in general
they are the most referred works in the literature. On
the other hand, the DISSECT-CF simulator is proven
to be much faster, scalable and reliable then CloudSim
(see (Mann, 2018)). This former research showed that
the simulation time of DISSECT-CF is 2800 times
faster than the CloudSim simulator for similar cloud
use cases, therefore we have chosen to analyse their
CLOSER 2020 - 10th International Conference on Cloud Computing and Services Science
194