Towards cloud based mobile ad hoc network simulation

Noman Islam
Indus University, Pakistan
[email protected]

Zubair A. Shaikh
National University of Computer and Emerging Sciences, Pakistan
[email protected]

Ghazala Shafi Sheikh

Indus University, Pakistan
[email protected]

Abstract: Simulations are considered as the most convenient A large number of simulation software has been proposed
approach for validating any research proposal. A simulator in literature [1, 2]. Among them, NS-2, Opnet and JIST/SWANS
attempts to model a real-world setting in order to analyze are some of the popular software. These simulation tools
the behavior of any algorithm. A large number of simulation generally provide a scripting language, such as TCL, for
software for ad hoc network has been proposed in literature. modeling simulation scenarios. The details of the various entities
These simulators are based on execution of simulation on involved in simulation are written in programming languages
standalone computers. This paper introduces the notion of like C/C++ and Java. In addition, a number of statistical and
simulation as a service for mobile ad hoc networks visualization packages are often supported by the software. The
simulation based on cloud computing. Cloud computing is an problem with current simulation software is that most of them
emerging discipline that enables outsourcing computing task are not suitable for simulating large scenarios. Some efforts have
remotely based on pay-as-you-go model. The paper provides been put in this direction to produce scalable simulation tool kits
a generalized architecture to perform ad hoc network (see details in section 2). JIST/SWANS, for instance, provides
simulations on cloud platforms. Similar to different service scalability by embedding simulation time into a virtual machine
models provided by cloud, some service models have been [3].
introduced for cloud based simulation. The proposed
approach provides different advantages such as dynamic This paper presents a novel cloud based approach to
service provisioning to support large scale simulation, simulation. A novel tool called Online Ad hoc Network
sharing of simulation models and reduced cost for simulation Simulation System (OASIS) has been proposed. According to
etc. The proposed tool has been implemented on Heroku Cayirci [4], Modeling and Simulation as a service is a model
platform that demonstrates the scalability of proposed for provisioning modeling and simulation (M&S) services on
approach. demand from a cloud service provider (CSP), which keeps the
underlying infrastructure, platform and software
Keywords: ad hoc network, simulation as a service, cloud based requirements/details hidden from the users. Cloud computing
simulation, collaborative simulation has emerged as a new model of computing. It comprises a set of
pooled resources available to users as a utility. This enables
1 Introduction execution of high performance computing (HPC) algorithms.
The paper presents a generalized architecture to simulation based
Simulation has been considered as amongst few effective
on cloud computing. The approach defines different types of
strategies for validating any research proposal. They are
service models similar to service models supported by
especially useful in situations where the environment imposes
conventional cloud computing systems.
different constraints in implementation of a proposition. Ad hoc
network (an infrastructure-less wireless network) is one such
Rest of the paper discusses the proposed scheme. In the
area where it is often impossible to perform real-time
next section, a review of literature relevant to cloud based
implementation of any proposed solution.
simulation is provided. The proposed approach to cloud based
simulation is then presented. The discussions on implementation
A testbed provides an alternative to analyze and validate a
and evaluation of proposed approach are then provided. The
research proposal in lab-settings. However, testbeds are not
paper concludes with discussions on limitations and possible
economically feasible. In addition, it is very difficult to deploy,
extensions of proposed work.
control and monitor a testbed environment. Simulations are an
alternative and cost effective approach for the validation of
scientific proposals. A simulator is a piece of software that 2 Literature Review
mimics the behavior of any real world environment. It provides a In literature, a large number of software has been proposed
handy approach to conduct research experiments. The input to for simulation of MANET. Some of these tools are NS-2,
any simulation software is a set of scenarios and the outputs are Glmosim, Opnet, JSim, OMNet++, QualNet and JIST/SWANS
usually captured in trace file. The input scenarios are [1, 2, 5]. Among these tools, only few provide support for large
reproducible thus enhances the credibility of the experiments. scale simulation. For instance, Glomosim [6] enables scalable
The trace files can be used for further analysis. simulation using parallel discrete simulation capability provided
by Parsec and can be run on shared memory symmetric The simulations can be accessed globally irrespective of the
processor computer. JIST/SWANS [7] is a Java based location. Even the simulation can be controlled from thin clients
simulation tool that provides a highly efficient sequential like mobile devices. Finally, the cloud can be easily
simulation engine. It adopts virtual machine based simulation i.e. administered, can be deployed with very low deployment cost,
simulation time advances not by any clock but entities and requires very low associated cost for maintenance of
themselves. The software claims to simulate networks twice as hardware.
large as what can be simulated by NS-2 and GlmoSim.
3.2 Proposed architecture
There have several efforts put in literature to provide Figure 1 shows the layered architecture of proposed cloud
modeling and simulation as a service on cloud platform. In [8], based simulation. At the infrastructure layer resides different
the concept of modeling and simulation as a service is types of computation and storage resources that are dynamically
introduced. In [9], the authors presented a polymer portal which provisioned using the hypervisor software. The platform layer
is motivated by the fact that simulation software are often hosts the actual operation system. Then, at the simulation layer,
expensive to use. So, a user can learn the simulation software various components of simulation (as discussed later on) are
first before actually buying the simulator. In addition, the available as services. These services can be accessed using
simulation can be rented when needed to reduce the cost. The application interfaces or directly by client.
authors in [10] proposed that all the activities of modeling and
simulation ranging from development and execution to analysis
should be performed online via a web browser on a cloud
platform. Cayirci [4] presented a survey on cloud based
modeling and simulation. In [11], Simulation Software-as-a-
Service (SimSaaS) is proposed for cloud that provides rapid
simulation in an elastic cloud environment. In [12], the authors
presented a network simulation environment based on the
concept of cloud computing.

Based on the literature review, it is found that there has not

been any effort put in to simulate ad hoc networks on cloud
platform and the work presented in this paper can be considered
the first proposal in this direction.

3 Cloud based simulation of mobile ad

hoc network
This section presents the proposed approach to cloud based
simulation. The basic motivation behind the proposition is the Figure 1: Proposed architecture for cloud based simulation
work on gird portal proposed for solving scientific applications
[13]. Grid computing combines a large number of loosely Figure 2 shows the various components of proposed
coupled resources belonging to different administrative domains simulator. Every virtual machine hosts a simulation instance for
for solving computationally intensive problems. Grid portals an end user. A simulator comprises simulation engine,
provide a transparent, one point access to grid resources for simulation library, simulation scripts and traces of simulation.
solving scientific problems. It is believed that similar to grid Different user interfaces are provided to configure and control,
portals, an interface to cloud environment for submission of jobs visualize and analyze any simulation. In addition, scripts can be
and simulation of different research algorithms can be deployed directly by end user interacting with simulator.
developed. This provides a much more simplified approach to
large scale simulation. 3.3 User interfaces
The user interface provides an intuitive interface to end user
3.1 Advantages of cloud based simulation for interaction and creation of simulation networks. Figure 3
The development of a cloud based simulator provides shows the snapshots. The control panel (Figure 3a) provides an
several key benefits. The first and foremost important benefit is interface for the management of various designing options.
the dynamic provisioning of resources. The computational
resources can be dynamically allocated for simulation when ever The control panel provides options to:
required. This is unlike the conventional simulation approach Add a New Node in the current configuration
that will not work for very large number of nodes. Load a previously saved configuration
Save current configuration
The second advantage is that the simulation models are Visualize the network
accessible to wide range of researchers. Hence, experiments and View Details of a node
simulation scenarios can be reproduced by researchers to analyze Start the simulation
and extend any approach proposed by other researchers. Another
advantage of cloud based simulation is collaborative research
and sharing of important results.
 3.4 Simulator
The simulator is responsible for execution of scripts and
delivering the output of the simulation results to the end user.
The client is relieved of the processing capabilities and the
execution of the simulation scripts are handled at cloud server.
So the client requirements are minimized and the simulations are
optimally executed at the server end. The results of the
simulation are saved as a trace file and can be visualized on any
well known visualizer like NAM [14].

3.5 Service Models

The proposed approach exposes different types of
simulation services to end users as discussed below.

3.5.1 Configuration-as-a-Service (CaaS):

Using this service, a simulation scenario can be configured.
A user can add/ delete a node, view/ set properties of node, load/
save a configuration. These configurations are saved on the
cloud.
Figure 2: Components of proposed simulator
3.5.2 Deployment-as-a-Service (DaaS):
The Add New Node option provides the interface to Simulation scripts can be deployed at cloud servers using
configure the new node and starts the execution of node etc. this service. Once a simulation script is deployed, it can be
Figure 3b shows the screen for configuration of a new node. It executed on cloud. The simulation scripts and results are also
provides configuration for node identification, nodes IP address accessible to a wide range of researchers for reproducing the
and options for creating and removing links with other nodes etc. simulation experiments.
The visualizer is shown in Figure 3c. The nodes are represented
as circles with their colors representing various states (such as 3.5.3 Simulation-as-a-Service (SimaaS):
sending a message and receiving a message). A configured and deployed simulation can be executed
using this service. The simulation is executed and results are
saved on cloud storage which is then accessible to end user and
community of researchers.

3.5.4 Visualization-as-a-Service (VaaS):

One a simulation is running, it can be visualized in runtime
using this service. Figure 3c shows snapshot of visualization

4 Implementation Details
The proposed approach has been implemented on Heroku
cloud platform [15]. As a bench mark algorithm, the service
a) Control Panel b) Configuration Panel discovery scheme presented in [16] has been implemented. The
relevant scripts and different simulation scenarios have been
executed. Figure 4 shows the snapshot of simulation. The
various messages exchanged among the nodes, logs and service
records are shown in figure. The simulator implements various
(such as MAC, transport and routing) protocols, statistical
analysis tools and algorithms for ad hoc network simulation.
Two basic routing protocols i.e. AODV [17] and DSR [18] are
provided for simulation. The simulation results are generated as
trace files that can be read by any visualizer such as NAM. The
proposed approach has been analyzed by varying the number of
nodes as well as services in the network.

c) Visualizer
Figure 3: User interfaces for cloud based simulation
 Figure 6 shows the result with varying number of
services. It can be seen that as the number of service increases,
the cloud based simulation shows linear degradation in
performance while the non-cloud simulation degrades very
sharply. The results confirmed that the cloud based simulation
environment is scalable and is able to be simulated virtually for
any resource requirements.

70
60

Unanswered Requests
50
40
Cloud based
30 Simulation
20 Non Cloud
Simulation
10
Figure 4: Snapshot of simulation 0
10 50 100 200 500
Services
Figure 5 shows the performance of our simulator when it
is run on cloud computing environment of Heroku. The results Simulation Parameters
are compared with the same simulation run on a non-cloud
environment. The simulations were performed with different Simulation Time 300 sec
number of nodes. The results are obtained after 300sec of Nodes 100
simulation. Hence, some service discovery requests remain Service Advertisement Frequency 10 / sec
unanswered. From the graph, it can be seen that with the rise in
number of nodes, the number of service discovery requests that Requests Issued 30 / min
remain unanswered rises very sharply in non-cloud based
simulation. While in the cloud-based simulation, the rise in Figure 6: Performance with varying number of services
unanswered discovery requests is linear. The reason is that the
cloud based simulation dynamically provisions itself with the 5 Open research issues in cloud based
rise in need of resource requirements. Hence, performance
doesnt degraded sharply.
simulation
Cloud based simulation is a new paradigm that opens
various research issues. This section outlines some of these
60
research issues and limitations of cloud based simulation. The
50 first challenge is the emergence of new security threats. The
Unanswered Requests

simulation performed on the cloud servers can be exposed to

40 privacy issues. The datasets used for simulation can be of
proprietary nature. Insecure interfaces and APIs, weakness of
30 Cloud based
Simulation hypervisors and account hijacking can be exploited to
compromise the privacy and integrity of cloud data. The
20 Non Cloud architecture of cloud (public, private or community cloud)
Simulation
10 becomes an important question in this situation. The availability
of services becomes an issue as cloud services can become an
0 attractive point for cyber attacks. Another important research
10 100 200 500 1000 issue is the service model supported by the cloud. What are the
Nodes various types of simulation services supported and how many
such services are sufficient?
Simulation Parameters
Simulation Time 300 sec 6 Conclusion
Services 50 This paper proposes a cloud based and scalable approach to
Service Advertisement Frequency 10 / sec ad hoc network simulation. A generalized architecture has been
Requests Issued 30 / min proposed and implemented, and different types of services have
been identified. Different scenarios for evaluation of proposed
Figure 5: Performance with varying number of nodes approach have been discussed. The proposed tool provides a
novel approach that can be used in combination with other
techniques e.g. parallelization techniques adopted by other
simulators. New types of service models can be introduced for
cloud based simulation. The proposed tool provides a reference
implementation of cloud based simulation. In future, different VINT network animator,," Computer, vol. 33, pp. 63-
simulation packages for implementation of MAC layer models, 68, 2000.
routing protocols, simulation visualizes, mobility models and [15] N. Middleton and R. Schneeman, Heroku: Up and
supporting tools for statistical analysis can be developed. Running: Oreilly, 2013.
[16] N. Islam and Z. A. Shaikh, "A novel approach to
service discovery in mobile adhoc network," presented
References at INCC, LUMS, 2009.
[17] "AODV Routing " Internet Engineering Task RFC
[1] D. Cavin, Y. Sasson, and A. Schiper, "On the accuracy 3561, 2003.
of MANET simulators," presented at Second ACM [18] "Dynamic Source Routing " Internet Engineering Task
international workshop on Principles of mobile Force RFC 4728, 2007.
computing, 2002.
[2] M. M. Koksal, "A Survey of Network Simulators
Supporting Wireless Networks," Middle East Technical
University Ankara 2008.
[3] R. Barr, Z. J. Haas, and R. v. Renesse, "JIST:
Embedding simulation time into a virtual machine,"
presented at EuroSim Congress on Modelling and
Simulation, 2004.
[4] E. Cayirci, "Modeling and simulation as a cloud
service: a survey," presented at Winter Simulation
Conference, 2013.
[5] T. Issariyakul and E. Hossain, Introduction to network
simulator NS2: Springer, 2011.
[6] X. Zeng, R. Bagrodia, and M. Gerla, "GloMoSim: a
library for parallel simulation of large-scale wireless
networks," presented at Twelfth Workshop on Parallel
and Distributed Simulation, 1998. PADS 98, 1998.
[7] F. Kargl and E. Schoch, "Simulation of MANETs: a
qualitative comparison between JiST/SWANS and ns-
2," presented at 1st international workshop on System
evaluation for mobile platforms, 2007.
[8] E. Cayirci, "Modeling and simulation as a cloud
service: a survey," presented at Winter Simulation
Conference: Simulation: Making Decisions in a
Complex World, 2013.
[9] T. Bitterman, D. Cai, D. Hudak, R. Ramnath, J.
Ramanathan, and B. Zhang, "Simulation as a Service: A
Cloud-Based Framework to Support the Educational
Use of Scientific Software," presented at 1st
International IBM Cloud Academy Conference, 2012.
[10] M. Seablom, "Modeling & Simulation as a Service
(M&SaaS)," NASA 2011.
[11] W.-T. Tsai, W. Li, H. Sarjoughian, and Q. Shao,
"SimSaaS: simulation software-as-a-service," presented
at 44th Annual Simulation Symposium, 2011.
[12] L. Bo-hu, C. Xu-dong, H. Bao-cun, L. Tan, Z. Ya-bin,
Y. Hai-yan, H. Jun, D. Yan-qiang, H. Ji-jie, S. Chang-
feng, T. Zhen, W. Peng, S. Guo-qiang, and W. Xiao-
hua, "Networked Modeling & Simulation Platform
Based on Concept of Cloud ComputingCloud
Simulation Platform," Journal of System Simulation,
vol. 17, 2009.
[13] M. P. Thomas, J. Burruss, L. Cinquini, G. Fox, D.
Gannon, L. Gilbert6, G. v. Laszewski, K. Jackson, D.
Middleton, R. Moore, M. Pierce, B. Plale, A.Rajasekar,
R. Regno, E. Roberts, D. Schissel, A. Seth, and W.
Schroeder, "Grid Portal Architecturesfor Scientific
Applications."
[14] D. Estrin, M. Handley, J. Heidemann, S. McCanne, Y.
Xu, and H. Yu, "Network visualization with NAM, the