A CLOUD BASED ARCHITECTURE FOR WORKING ON BIG DATA WITH WORKFLOW MANAGEMENT

International Journal of Web & Semantic Technology (IJWesT) Vol.9, No.2, April 2018
DOI : 10.5121/ijwest.2018.9201 1
A CLOUD BASED ARCHITECTURE FOR WORKING ON
BIG DATA WITH WORKFLOW MANAGEMENT
Tahereh Koohi-Var1
and Morteza Zahedi2
1
International Campus of Kharazmi, Shahrood University of Technology, Shahrood, Iran
2
CE Department, Shahrood University of Technology, Shahrood, Iran
ABSTRACT
In real environment there is a collection of many noisy and vague data, called Big Data. On the other hand,
to work on the data middleware have been developed and is now very widely used. The challenge of
working on Big Data is its processing and management. Here, integrated management system is required
to provide a solution for integrating data from multiple sensors and maximize the target success. This is in
situation that the system has constant time constrains for processing, and real-time decision-making
processes. A reliable data fusion model must meet this requirement and steadily let the user monitor data
stream. With widespread using of workflow interfaces, this requirement can be addressed. But, the work
with Big Data is also challenging. We provide a multi-agent cloud-based architecture for a higher vision to
solve this problem. This architecture provides the ability to Big Data Fusion using a workflow management
interface. The proposed system is capable of self-repair in the presence of risks and its risk is low.
KEYWORDS
Workflow Management Interface, Multi-Sensor Data Fusion, Big Data, Cloud Computing
1. INTRODUCTION
Managers often reuse and refine workflows or current patterns of processes [1, 2]. They can share
their workflows or create new workflows from a public repository or a new project space (the
extracted data, new components, sub-workflows, etc). Meanwhile, the use and fusion of
structured and unstructured datum that has a huge volume (Big Date [3]) and it seems to be a hard
work. On the other hand, if the process is automated by workflows, data sources are selected
during the run of workflow and the parameters are set by the user. Workflows may have resource
scheduling in High Performance Computing (HPC [4]) (for example in local computer clusters),
or time scheduling including remote resources (Grid computing [5] or Cloud [6]). Data may also
be stepped i.e., from certain places waiting for data, where computational tasks are done on HPC
clusters. For an instance of the application, it is important for marine vehicle to have tracking [7]
and navigation [8]. This tracking and navigation is needed along with other critical time and
inhibitor works in management. The nature of the inhibitor and critical time works has led marine
vehicle management systems towards systems with “Multi-INT” multi sensor data fusion (Figure
1). But, generally the design of these systems has challenges, because gaining reliable
information is difficult because of the complexity of the decision making in dynamic
environments. In fact, the designer of management systems have equipped management systems
with the ability of automated cognitive processes, on the importance of collecting various data
from different sensors. This equipment is for increasing the security and efficiency of marine
vehicles. In the equipped vehicles with the purpose of combining data and performing tasks as
workflows is to bring the marine vehicle to a situation to increase its ability. Actually, our
purpose in the management system of a marine vehicle is providing an interface for the tools that
are presented as heterogeneous agents for safe navigation and mission execution. The system
interfaces with various sensors to give real-time understanding of the situational maritime picture
to the operator (the user). These sensors can include: GPS/DGPS, Radar (S&X band, and TWS) ,

2
Compass, weather, Electronic Support Measurement (ESM), Precision Direction Finder (PDF),
RF system, Specific Emitter Identification (SEI) , Automatic Identification System (AIS).
The data set is supposed to help manage user-managed processes or controllers. In implementing
a management system this data and other structured data are involved, and all these must be
considered.
Figure 1 – Multi sensor data fusion of a marine vehicle
Since the marine vehicle has the reorganization ability based on a wide range of surface and sub-
surface sensors, the data may have come from a variety of sources so that the aggregated data
maybe vague and noisy. Here, the management system is required to provide a solution to the
data defect and maximize the target's success. This system is needed to be able to manage
processes, interpret and fuse the pieces of different types of information quickly. The dependency
between the data fusion, the sensor set, the goals and the environment make the implementation
of such a framework difficult. In marine environments, particularly, the large volume of work and
the complex signal emission environment, especially for observing underwater sonar, increases
the complexity of data fusion operations. The common feature of those systems that support data
fusion is that they have several sensors that operate with a variety of data sources, have limited
constant time for processes, and the decision processes are done real-time based on aggregated
information. A reliable data fusion model should meet this requirement and allow the user to
monitor data stream continuously. For this purpose, we propose a system for data fusion based on
the multi-agent system model presented in [13], using the real-time operating system, and an
Application Interface (API). This system must be capable to run the user-defined workflow
automatically and can handle Big Data. For this purpose, a model for managing the workflows of
this cloud-based system is presented that have the potential to manage risks. Finally, we evaluated
the presented system by simulating it in the CloudSim framework [22].
The structure of this paper is as follows: In section 2 we discuss about the works related to issues,
including issues related to middleware and workflow management system. Then in Section 3 we
will introduce the proposed architecture provided for data fusion. In Section 4, the model for
workflow management is provided based on cloud computing, and then we will provide a way to
improve it. In Section 5 we show the results and simulation. Finally, in Section 6 the paper is
concluded.

3
2. RELATED WORK
This study is concerned about the design of data fusion system, the middleware, the workflow
management system, and the proposed architecture. In the next sub-sections, we study each of
them.
2.1. DATA FUSION SYSTEM DESIGN ISSUES
Sloman [26] has a strong evidence for thinking and designing models at the level of architecture,
instead of building independent component units. Designing the real world data fusion requires
the collaboration of the distributed system (network), organizational concepts (command), and
environmental perception (textures). In the design of a data fusion system for a particular
composition task, the designer must choose from many selections are involved with trade-off.
Lambert [16] identified eight issues should be considered by the designers of the data fusion
systems within the scope of the Object assessment, Situation and Impact. The eight tradeoffs
illustrated in [16] are as follows:
1. Automation- from totally manual to fully automatic, with all levels between.
2. Aggregation- including communication topology, protocols, roles, and
bandwidth, etc.
3. Integration model - for the aspects of Human Machine Integration (HMI),
Machine-Machine Integration (MMI), and Human-Human Integration (HMI)
4. Situational awareness - by human, machine, or both.
5. C2 Fusion Policy - centralized versus decentralized.
6. Representation - expressivity versus calculability.
7. Information processing - analytical versus heuristic.
8. Integration of the machine.
Each of these tradeoffs that are well-distinguished by Lambert, requires a branch of extensive
research and expertise related to it. For example, consider the integration trade-off, which
includes networking issues and communication protocols. In the navigation management with the
security issue some protocols are required to be selected. The protocols may be capable to handle
cryptographic issues as well as network configuration policies. They must be taken so that the
datum is transmitted correctly. Meanwhile, a good data fusion system in the management is the
system that can meet the system's control needs independent of these issues, without worrying
about infrastructure technology. Therefore, in this paper, we will design an interface for workflow
management according to the following:
1. Lightweight: This interface needs to be implemented on low-level platforms.
2. Multi-agent: Processes are done by a collection of processing agents [13]. The
only requirement is that these systems are based on a real-time operating system
[18], or a middleware is used that provides the processing of real-time processes.
We'll explain this in more detail in the next section.
3. Portability: Data fusion processing platforms is not homogeneous (for example,
they can be Personal Digital Assistants (PDAs), Laptops or sensor nodes). The
data fusion interface is required to be not sensitive to the diversity of the operating
systems and the architecture between the network nodes.
4. Development: Modular construction ensures to make programming easy.
a. Scalability: The design requires scalability to be considered when increasing
network size.
The migration or distribution environment of our proposed architecture contains a key component
called middleware. In the reminder of this section we introduce the related works about this
component.

4
2.2. MIDDLEWARE
The most common middleware used in management systems is the web services [17] and
Common Object Request Broker Architecture (CORBA) [28]. But these have disadvantages. For
example, unlike the success of CORBA in the deployment of the basic architecture of the
distributed object systems, its learning is difficult, it is difficult to use, it suffers from several
designs, it has low performance in the protocol, and it is necessary to support some of the features
that always are required.
In this paper, we intend to consider our system based on multi-agent middleware, which benefits
from the support of dynamic configuration qualities, displaying knowledge and understanding the
relationships between elements involved. This middleware is required to provide different
possibilities depending on the type of heterogeneous problems, for realizing communications
between nodes which are distributed in a computational environment. Heterogeneity is often
restrained by standard communication protocols that is expected be involved with all software,
and technology network systems. To overcome the heterogeneity problems, we will provide a
framework based on the current multi-agent platform. For this purpose, we will examine the
current multi-agent platforms.
Current multi-agent platforms: JADE (Java Agent Development) is the most matured framework,
after more than ten years of development and is the reference of multi-agent system development
agrees with FIPA. Meanwhile, there are other related agent platforms that help developing these
types of tools. The most related of these tools are summarized in Table 1.
Table 1: Current related agent-based frameworks
Related
article
Name Language Standard Free
[29] JADE Java FIPA Yes
[9] Jadex Java FIPA Yes
[10] Cougaar Java No Yes
[11]
Agent
factory
Java FIPA Yes
[27] JAMES II Java No Yes
[12] JACK Java No No
[27] AnyLogic Java No No
Recently, a modern approach has been proposed for multi-agent development [27] that is
comparable with JADE, and has features such as the ability to implement the basic
management agent in C ++, Java and Python. This middleware is scalable, and it has many
agents and short response time. So, it is suitable for the data fusion interface that we will
provide.
2.3. ARCHITECTURE
Using a distributed system [25] allows the system processes to be managed at runtime. Among
the previous work that has been done to manage processes at the operating system level we can
refer to the work at [15]. The architecture presented in [15] is a distributed system. This system is
based on the pattern of service process communication via message transactions. The
communication protocol used in the system that proposed in [15] is TCP/IP and a system control
and monitor processor selects data in HTTP. In [15] each fused data has a link to the data
distribution system. The disadvantage of a distributed system over the multi-agent system is that
the elements of a distributed system cannot have the autonomy of agents in multi-agent systems
[13, 27]. Therefore, in this work, we propose a multi-agent based architecture that let the elements
act as autonomous.

5
2.4. WORKFLOW MANAGEMENT SYSTEM
Workflows are rule-based management programs that they model the execution of tasks to the
complex functions and direct the work [19]. Workflows are usually used to automate the hidden
layers from the user's perspective. Middleware like CORBA can also benefit from the workflows
for doing their work [19]. But for workflow management it is necessary to consider the volume of
data has become big, and so the management should be capable to work well with Big Data [4],
[6]. On the one hand, this system should be responsive to the big volume of the data and, on the
other hand, it should have low cost. The issue of lowering the energy costs has been investigated
in different ways in the system [24]. We therefore propose a method based on the cloud
computing that its cost is low as the services are provided over time.
3. ARCHITECTURE
The proposed architecture is constructed of radar, tracking fusion and workflow monitoring and
control agents. This architecture is a similar the architecture proposed in [15], with the difference
that in this architecture the multi-agent middleware is used. In this architecture, the user interface
and mission goals must be defined in the system control and monitoring agent. Figure 3 illustrates
the architecture proposed in this study. In this architecture, a multi-agent middleware is
considered, which is connected to 2 weeks and 24 hours storages. On the other hand, the
middleware is connected to acoustic sensor agent, data fusion agent, server, receiver and the
information agent. Each of these can work autonomously. Figure 2 illustrates the proposed
middleware based on [13]. The required middleware has a multi-agent class diagram.
Track Fusion
Agent
Acoustics Agent Information Agent
Agent
Management
System
Message
Transport Service
Agent
Directory
Facilitator
Comm
FIPA-Related Classes
Marine Navigation System
Related Classes
Resource
Management
Agent
Figure 2- Proposed Multi-agent class Diagram (Middleware)
The used communication protocol in the proposed architecture (Figure 3) is based on FIPA. This
protocol is usually used in the multi-agent systems and facilitated the development of the agent-
based applications [13].

6
Figure 3 – Proposed architecture schematic
The information agent is designed to monitor databases, raw sensor data filtering, topology
management, and so on. This agent takes the observations of the real world system. During the
execution of the workflow by the agent, the input datum has been consumed and new data
products have been created (in the track data fusion agent). In this architecture for implementing
large-scale performance (the workflow is played on hundreds or thousands of nodes: for hours,
days, or weeks at a time), real-time monitoring is also important: mid-data products, and specific
information about web-based monitoring is required to inform the user about the progress and
possible issues during execution.
Further details of this architecture are as follows:
Sensor Track Data Fusion: Figure 4 shows our proposed architecture. Sensor data (e.g.
ELINT/Radar, etc.) will be fused in the data fusion agent. In fact, the input is a real-world
observation to estimate the real world state. A weather prediction system, which is covered as an
agent, will be into this category. The output of this agent is based on the National Marine Electron
Standards (NMEA).
Track fusion agent: Modelling the navigational dynamics of the marine vehicle leads to 6 Degrees
Of Freedom (6DOF), so 3 orthogonal angle rates and 3 orthogonal accelerations are required.
Resource Management Agent (RMA): In this architecture, we separate information fusion
functions and Resource Management (RM). This separation was the goal of the promotion from
the JDL models to DFIG [14]. RM is a set of management tools, including the management
sensors, workflows and manoeuvre, platform placement and the user selection for meeting the
mission goals. Since the unobserved aspects of situation awareness problem cannot be processed
by a computer, the user's knowledge, and reasoning are essential; in fact, to provide and maintain
the Quality of Service (QoS), RM must be guided by QoS. The workflow management is done by
this agent.

7
Figure 4. Fusion Architecture core Schematic
Figure 5 shows that at the core of the proposed architecture the RMA agent is consisted from the
manoeuvre management, sensors and workflow agents.
Figure 5. Interface between sensor management, workflow, and manoeuvre agents
RMA is an estimate of the current state of the global situation and a plan or a produced solution.
Such agent can use other agents to for example find additional information. In the next section,
we explain the workflow management model to implement by this agent.
4. WORKFLOW MANAGEMENT MODEL BASED ON CLOUD COMPUTING
To work with Big Data, a cloud-based management model is selected for workflow management.
With the transfer of the data in the cloud, the cloud computing has the challenge of security and
the security measure is unknown [21]. The distance between the user and the physical location of
the data creates a barrier called risk. Risk is the possibility of accident happening and has an
unwanted influence on performance goals. Different types of security risks, integration,
accessibility, and efficiency are raised. The risk characteristics of the impact, likelihood, and
effects for categories of the types of strategic, technical, operational, and rule based. These are the
bases of the work: the risk management problem formulation, risk modelling, risks scoring,
flexible risk system modelling, and finding control point of the cloud computing, modelling to
achieve decision making response, and providing the optimal solution for decision making. In this
regard proposing a model for information risks assessment is necessary to assess the allocation

8
and maintenance of resources. The following methodology is our proposed approach for
evaluating the efficiency of workflow management system in the presence of risk.
System performance improvement in the presence of risk: the proposed approach to risk
management in workflows is based on the work presented in [23]. We initially estimate the
system failure rate based on the Poisson distribution like [24]. Then, we must balance the systems
loads with the migration of workflows in the proposed framework. In fact, migration is a basic
concept of the mechanism of proposed workflow management method. Thus, we give the cloud
system the capability of self repair to have more safety against risks. In the next section, we
evaluate our simulations regarding the proposed framework.
5. SIMULATION AND EVALUATION
For simulation we have used Cloudsim [22]. First, we proposed a cloud platform based on the
characteristics of Table 2 and we considered having per physical machine a virtual machine. In
fact, each host in the cloud is dedicated to a virtual machine. Then, we calculated the system risks
based on the failure rate that may be caused by the attacks on the cloud.
Table 2: Cloud specifications
Virtual machine operating system Processor RAM Workflows
50 Linux 4 cores 4 2
In this model, the failure rates are applied randomly with a Poisson rate to a range of workflow
(Cloudlet). In some runs, we may not have failures. But, the system in the case of failure can
repair itself by reallocating the virtual machine list to the system hosts.
Figure 6. Cloud repair in the workflow management system-vertical axis: success rate, and the horizontal
axis: Time
Figure 6 shows the system ability in self-repair. It shows that the success rate is increased over
time that it shows the system has the ability of self-repair in the case of failure. Figure 7 shows
the availability rate of the system in the same conditions.

9
Figure 7. Workflow management system availability to the cloud penalty- vertical axis: penalty, and the
horizontal axis: availability
Figure 7 shows the system has the ability of self-repair. The results form Figures 6 and 7 show
that the cloud risk is low.
6. CONCLUSION
In this paper, we presented an interface for workflow management to monitor the
performance of a marine, and then we stated its challenges. We then presented an
architecture that can handle the process management needs in the case of navigation and
tracking, according to today technology. This architecture can meet Big Data
requirements in implementation. On the one hand, because it is a cloud-based workflow
management model and on the other hand the results show that it can work well in the
presence of risk. Migration has the main role in this model in the self-repair and the
system control in the presence of risk.
ACKNOWLEDGEMENTS
The authors wish to thank their colleagues at Web Mining and Pattern Recognition (WMPR) Lab
of Shahrood University of Technology for their participation, insights, and patience.
REFERENCES
[1] T. Koohi, and M. Zahedi, Scientific Workflow Clustering based On Motif Discovery, International
Journal of Computer Science, Engineering and Information Technology (IJCSEIT), 7 (4), 2017.
[2] T. Koohi, and M. Zahedi, Linear Merging Reduction: A Workflow Diagram Simplification Method,
8th International Conference on Information and Knowledge Technology, 2016.
[3] F. Bajaber, R. Elshawi, O. Batarfi, A. Altalhi, A. Barnawi, Sh. Sacr, Big Data 2.0 Processing
Systems: Taxonomy and Open Challenges, Journal of Grid Computing, 14 (3): 379-405, 2016.
[4] E. Deelman, T. Peterka, I. Altintas, Ch.d. Carothers, KK van Dam, K. Moreland, M. Parashar, L.
Ramakrishnan, M. Taufer, and J. Vetter, The Future of Scientific Workflows, International Journal of
High Performance Computing Applications, p. 1-17, 2017.
[5] AG Finogeev, DS Parygin, AA Finogeev, The convergence computing model for large sensor data
mining and knowledge discovery, Human-centric Computing and Information Sciences, 7 (1): 11,
2017.

10
[6] P. Kacsuk, J. Kovacs, Z. Farkas, The Flowbster Cloud-Oriented Workflow System to Process Large
Scientific Data Sets, Journal of Grid Computing, 1-29, 2018.
[7] T. Koohi, F.K. Dehkordi, A. Izadi-Pour, and M.N. Fesharaki, Optimization of Multi-Target Tracking
in a Multi-Agent Architecture with Multi-Sensor Data Fusion, Majlesi Journal of Electrical
Engineering, 4 (1): 2015.
[8] T. Koohi, M. Yaghoobi, and S. Toosizadeh, Advances in Fuzzy Path Following of a WMR to
Guarantee Successful Navigation, 5th International symposium of SASTech, 2011.
[9] A. Pokahr, L. Braubach, W. Lamersdorf, Jadex: A BDI reasoning engine. Multiagent Systems
Artificial Societies and Simulated Organizations, Springer, 15: 149-174, 2005.
[10] A. Helsinger, T. Wright, Cougaar: A robust configurable multi agent platform. IEEE Aerospace
Conference, 1-10, 2005.
[11] R. Ross, R. Collier, GMP OHare, AF-APL-bridging principles and practice in agent-oriented
languages. In: Second International Workshop on Programming Multi-Agent Systems (ProMAS'04).
Lecture Notes in Computer Science, Springer, 3346: 66-88, 2004.
[12] M. Winikoff, JACK TM intelligent agents: an industrial strength platform. Multiagent Systems,
Artificial Societies, and Simulated Organizations, Springer, 15: 175-193, 2006.
[13] D. Vallejo, J. Albusac , JA Mateos , C. Glez-Morcillo, L. Jimenez, A Modern Approach to Multi-
Agent Development, The Journal of Systems and Software 83: 467-484, 2010.
[14] E. Blasch and S. Plano, DFIG Level 5 (User Refinement) issues, Supporting Situational Awareness
Reasoning, In Proceedings of the International Conference on Information Fusion 2005.
[15] DJ Bielecki, NRL Global Vessel Tracking Project (VTP), Naval Research Laboratory, 2008.
[16] DA Lambert, Tradeoffs in the design of higher-level fusion systems, Proceedings of the International
Conference on Information Fusion, 2007.
[17] D. Booth, H. Haas et al. (Eds.) Web Services Architecture, W3C Working Draft, 11 February 2004.
http://www.w3.org/TR/2004/NOTE-ws-arch-20040211/
[18] KH Kim, C. Subbaraman , Principles of Constructing a Timeliness-Guaranteed Kernel and Time-
Triggered Message-Triggered Object Support Mechanisms, ISORC, pp. 80-89, 1998.
[19] F. Ranno, and SK Shrivastava, A review of distributed workflow management systems, in the
Proceedings of the international joint conference on Work activities coordination and cooperation,
1999.
[20] R. Wu, Y. Chen, E. Blasch, B. Liu, G. Chen, D. Shen, A container-based elastic cloud architecture for
real-time Full-Motion Video (FMV) target tracking, In Applied Imagery Pattern Recognition
Workshop (AIPR), pp. 1-8th of April 2014.
[21] M. D. Ryan, Cloud Computing Security: The Scientific Challenge, and a Survey of Solutions, Journal
of Systems and Software, 2013.
[22] RN Calheiros, R. Ranjan, A. Beloglazov, CA De Rose, R. Buyya, CloudSim: a toolkit for modeling
and simulation of cloud computing environments and assessment of resource provisioning algorithms.
Software: Practice and experience, 41 (1): 23-50, 2011.
[23] JO Fitó, J. Guitart, Business-driven management of infrastructure-level risks in Cloud providers,
Future Generation computer systems, 32: 41-53, 2014.
[24] K. Djemame, D. Armstrong, J. Guitart, M. Macias, A Risk Assessment Framework for Cloud
Computing, IEEE Transactions on Cloud Computing, 4 (3): 265-78, 2016.
[25] H. Pérez , JJ Gutiérrez , S. Peiró, A. Crespo, Distributed architecture for developing mixed-criticality
systems in multi-core platforms, Journal of Systems and Software, 123: 145-59, 2017.
[26] A. Sloman, What sort of architecture is required for a human-like agent? In Foundations of Rational
Agency, ed. M. Wooldridge, and A. Rao. Dordrecht, Netherlands: Kluwer Academic Publishers,
1999.

11
[27] K. Kravari, N. Bassiliades, A survey of agent platforms, Journal of Artificial Societies and Social
Simulation, 18 (1): 11, 2015.
[28] W. Henrich, Th. Kausch F. Opitz, Data Fusion for the Finnish Fast Attack Craft Squadron 2000:
Concept and Architecture, In the Proceedings of the International Conference on Information Fusion,
2004.
[29] FL Bellifemine, G. Caire, A. Poggi, G. Rimassa, JADE: A software framework for developing multi-
agent applications. Lessons Learned, Information and Software Technology, Elsevier 5 (1-2), 10-21,
2008.
AUTHORS
Tahereh Koohi-Var is a PhD student at WMPR Lab of Shahrood University of Technology with Master of
Science from Azad University of Mashhad (2012). She obtained Bachelor Degree in Computer Engineering
from Azad University of Mashhad in 2009. Her researches are in fields of data fusion, signal processing,
and business process management systems. Recently, she has focused on BPMSs. She is currently CEO at
enahang Co. Ltd.
Morteza Zahedi is an Assistant Professor at Shahrood University of Technology. He is currently CEO at
Ganjineh Co
.

A CLOUD BASED ARCHITECTURE FOR WORKING ON BIG DATA WITH WORKFLOW MANAGEMENT

More Related Content

What's hot (20)

Similar to A CLOUD BASED ARCHITECTURE FOR WORKING ON BIG DATA WITH WORKFLOW MANAGEMENT (20)

Recently uploaded (20)

A CLOUD BASED ARCHITECTURE FOR WORKING ON BIG DATA WITH WORKFLOW MANAGEMENT