A Scalable Cloud Infrastructure Architecture for Simulation

This report presents a cloud infrastructure architecture for distributed simulation, addressing the challenges of simulating complex models on cloud platforms. It discusses the research hypothesis, which aims to connect different simulation techniques to facilitate large-scale model analysis and replication in various cloud environments, ultimately improving performance, scalability, and cost-effectiveness. The methodology employs a case study approach, drawing data from the assembly line of Ford Automobile Company to evaluate the results of auto-scaling the cloud infrastructure for distributed simulation. The research highlights the significance of distributed simulations in high-performance computing, defense, and gaming, and concludes with the achievements and future aims of Ford in leveraging distributed simulation.

A Scalable Cloud Infrastructure Architecture
for Distributed Simulation
Author’s Name-
Abstract— This thesis consists of Cloud Infrastructure
Architecture for Distributed Simulation. In this report, the
problem that is solved while dealing with the distributed
simulation is clearly discussed. The research that is held for this
thesis is also mentioned and the corresponding research
hypothesis is discussed briefly regarding the research paper. The
methodology that is used in the thesis is a case study methodology
and the results that is achieved is solved in this report.
Keywords—cloud computing; distributed simulation; SaaS;
PaaS; IaaS; Ford Automobile
I. PROBLEM TO BE SOLVED
The problem that is to be solved in this research
paper is about a cloud infrastructure and then this paper
proposes a solution to all the problems that is attributed to a
single model that is simulated on particular cloud platform.
Cloud computing is basically a model that enables the network
access that are on demand for sharing a pool of resources [6].
The service models of cloud are mostly chosen with the type
of resource which is taken. The distributed simulation
involved with cloud computing helps to study the
methodologies and the techniques which are needed for
executing simulation models and are also used for defining the
simulation models on architectures of distributed computing
[2]. A distributed simulation is generally a program software
that are used for evolution of models for some of the abstract
or real system. To evaluate the systems even before they are
built, the process if simulation is used. Simulation also affects
the changes that are on the existing systems. The simulation
process mostly uses set of variables. The distributed
simulation depends on dividing the simulation model over
many execution units. The entire execution unit manages only
a single part of model. In simulation process, execution
process handles the local event list and also generates events
logically that are needed for delivery for executing the remote
units [7]. Simulation process allows all processor to control a
part of state space and small fraction of events. To integrate
the simulators, which are distributed geographically, are done
by distributed simulation.
II. RESEARCH HYPOTHESIS
The hypothesis of this research paper is that by the
end of this research paper, different techniques for simulation
are expected to be connected and then work together so that
they can facilitate the complex and large models that can be
run, analyzed in many cloud platforms, and replicated. The
hypothesis made in this paper helps to improve the
performance, scalability, and execution speed and also reduces
the cost of modeling [5]. The contributions that were made
while working on this thesis is the deployment of distributed
simulation on different cloud infrastructure for the testing on
user.
III. CONTRIBUTION OF RESEARCH
In today’s modern digitization era, the computing
architectures and the networking systems brings the system of
computer parallel to masses, which generally increases the
potential number of users for those kinds of system. Cloud
computing is a process that allows all the applications to scale
the resources that are offered from large pool. The cloud and
the multi - core processor systems requires applications that
are modified to get an advantage on features that are mainly
provided by them. In this technological era, the activities that
are included in the business are mainly computerized and to
access them from anywhere and at anytime is a way to remain
competitive to achieve their business goals. By the process of
cloud computing, the workflows of the businesses can be run
made smooth, enabling a high performance in their business
[4]. The technology of distributed simulations mainly gains
the significance mainly in three different communities. The
different communities are community of high performance,
community of defense and the gaming industry that mainly
focuses on time management. The beginning of distributed
simulation was in the year 2005. The standards of
interoperability were being investigated that were based on
format of data exchange and also the protocols for CSPs.
There were also use of emulators used for feasibility study and
are also used to demonstrate the benefits of the stakeholders
when the emulators were adopted.
IV. EVALUATING THE RESULTS
Auto-scaling of the cloud infrastructure helps to
allow the safe running of the distributed simulation on all the
models involved in supply chain and the manufacturing
industries with reliability and speed. This paper has stuffs

collected from many literature reviews, which describes the
PaaS, SaaS and IaaS infrastructure [1]. The methodology that
is involved in this research paper is a case study approach. The
data that are involved in this research are collected from the
assembly line of production. The program fragment of the
prototype will be developed and the distributed simulations of
test run for various replications are also implied in this
research paper.
V. RESULTS ACHIEVED
The case study that is highlighted in this research
paper is about Henry Ford of Ford Automobile Company who
aims to produce high speed cars at lower production cost [3].
The research paper also shows the future aim of the Ford
Automobile Company and the steps that are involved with the
company for distributed simulation.
REFERENCES
[1] D'Angelo, G., 2011. Parallel and distributed simulation from many cores
to the public cloud, Proceedings of the 2011 International Conference
on High Performance Computing and Simulation, HPCS 2011 2011, pp.
14-23.
[2] Taylor, S.J.E., BØhli, L., Wang, X., Tuner, S.J. and Ladbrook, J., 2005.
Investigating distributed simulation at the Ford motor
company, Proceedings - IEEE International Symposium on Distributed
Simulation and Real-Time Applications, DS-RT 2005, pp. 139-147.
[3] Ford Motor Company. 2017. 100 Years of the Moving Assembly Line.
[ONLINE] Available at: http://corporate.ford.com/innovation/100-years-
moving-assembly-line.html. [Accessed 22 November 2017].
[4] Anagnostou, A. and Taylor, S.J.E., 2017. A distributed simulation
methodological framework for OR/MS applications. Simulation
Modelling Practice and Theory, 70, pp. 101-119.
[5] Gabriele D’Angelo. Parallel and Distributed Simulation from Many
Cores to the Public Cloud. Proceedings of the 2011 International
Conference on High-Performance Computing and Simulation (HPCS
2011), Istanbul (Turkey), IEEE, July 2011. ISBN 978-1-61284-382-7.
[6] Soliman, H.M. and Elmaghraby, A.S., 1996. Efficient clustered
adaptive-risk technique for distributed simulation, IEEE International
Symposium on High Performance Distributed Computing,
Proceedings 1996, pp. 383-388.
[7] Fujimoto, R.M., 2001. Parallel and distributed simulation
systems, Winter Simulation Conference Proceedings 2001, pp. 147-157.