A SERVICE ORIENTED ENGINEERING APPROACH TO
ENHANCE THE DEVELOPMENT OF AUTOMATION AND
CONTROL SYSTEMS
David Hästbacka and Seppo Kuikka
Department of Automation Science and Engineering, Tampere University of Technology
P.O. Box 692, FIN-33101 Tampere, Finland
Keywords: Information services, Expert services, System design, Design process, Automation and control system
development, SOA.
Abstract: In order for the manufacturing industry and closely related engineering disciplines to be competitive and
productive, business structures and practices have to adapt to global changes and harder competition on all
levels of operation. An engineering approach based on engineering services provides a foundation for
commercial of the shelf services to be combined and utilized between engineering enterprises in the
development of automation and control systems. A service based operations model would enable the
utilization of expert services as a part of the development process to improve system quality, increase
productivity and provide better work process management as well as allow easier integration to later life-
cycle operations. This paper presents opportunities this kind of a conceptual approach offers and outlines
some of the related research challenges that need further investigation.
1 INTRODUCTION
Globalization and the transition to a post-industrial
society have had a significant impact on the
manufacturing industry and the business structure of
today. As a result from manufacturing facilities
moving to cheap labour countries many associated
businesses are under constant pressure to keep up
with increasing competition. This also affects related
engineering disciplines, such as automation and
control system design, which have to cope with
global competition and interweaving work processes
in a distributed network of business partners and
sub-contractors.
In order for these engineering businesses to be
competitive they have to increase their productivity
and be able to organize their services more flexibly
across organisational boundaries. In addition, the
quality of systems designed has to improve as
requirements on system properties such as
reliability, accuracy and maintainability are
tightening.
Business re-engineering and the restructuring of
business processes towards a service based
operations model brings agility and flexibility to
better adapt to changing circumstances and business
needs. An engineering process composed of
individual services can increase productivity by
providing easier management of involved
operations. Some of these services are commercial
solutions from external service providers and some
are redefined and restructured from existing in-
house activities. While many of the tasks in
development still require active human participation
there are possibilities to implement some of the
services as automated IT-based services. As a result,
the management of well-defined service units, their
compositions, and their interaction and orchestration
becomes easier to handle from a workflow point of
view.
From the business point of view services should
be seen as repeatable tasks that perform some
actions in order to complete a larger assignment.
Service-orientation, the next generation of business
automation, is a way of organizing and integrating
business processes from these services. It should be
noted that in this context services are considered as
parts of the business processes, and service-oriented
systems offers a way to provide existing functions
and enable new business process models.
219
Hästbacka D. and Kuikka S. (2009).
A SERVICE ORIENTED ENGINEERING APPROACH TO ENHANCE THE DEVELOPMENT OF AUTOMATION AND CONTROL SYSTEMS.
In Proceedings of the 11th International Conference on Enterprise Information Systems - Information Systems Analysis and Specification, pages
219-224
DOI: 10.5220/0002007702190224
Copyright
c
 SciTePress

Service-oriented architecture (SOA) is an
implementation-agnostic architecture for linking
activities and resources that each addresses a
specific task. From a technical perspective service-
oriented systems allow encapsulation and
componentization of functionality into reusable
services with well-defined specifications that can be
discovered, bound and invoked by service
consumers. At the same time an SOA approach
needs to consider business processes both in general
and with domain specific special characteristics in
mind. Therefore it is different and much more
complicated to provide engineering as a service
compared to services such as online ticket
reservations.
The engineering business needs to be viewed
within a service-oriented context to consider
additional requirements on current practices and
technological infrastructure. This paper introduces
some of the current practices and problems in
automation and control system design and highlights
related challenges during later life cycle phases.
Based on these presumptions, following is a
discussion on what could be enhanced and a
conceptual presentation of possible solutions based
on services and the sophisticated utilization of
information systems available.
2 RELATED WORK
Service-orientation has been a popular catchphrase
during this decade. A lot of research work has been
conducted and considering software engineering
especially in the fields of enterprise systems and
business intelligence. Service based engineering and
workflows have also been studied and according to
Bergman et al. (2002) future users will request
engineering services automatically from an
automated server in the same manner as stock quotes
and weather services are requested today. Bergman
et al. have studied the challenges of providing
engineering as services and developed an automated
workflow system prototype in the aerospace
telecommunication domain to support engineering
activities. They concluded that applications of this
type require considerable domain specific input,
have strong dependencies on the environment and
typically require a large number of parameters.
A service-oriented architecture is an architectural
approach of organizing software units focusing on
services and their interaction. The approach
considers both service provider and consumer needs
and enables the use of services in the business
model. For IT systems this means automatic
utilization of services as a part of the business
process. Lee et al. (2007, 2008) have studied the use
of services in product development in the
mechanical industry and developed an engineering
framework for composing services supporting QoS
attributes and methods typical to business process
management such as BPEL (Business Process
Execution Language).
Thramboulidis et al. (2008) have studied SOA-
based embedded system development for industrial
applications and utilized semantic web technologies
in their prototypes. In their research, they outlined a
conceptual model of a semantic web-based
framework containing services, such as the ones for
component and device repositories, to be embedded
in service-oriented development environments. They
also created prototypical Web services for embedded
system development environments and demonstrated
the applicability of the proposed approach.
Francois Jammes and Harm Smit (2005) have
studied service-oriented paradigms in industrial
automation and networking of intelligent devices.
They considered Web services as the enabling
technology for service-oriented applications to be
flexible and adaptive to changing strategies. Fuji and
Suda (2005) have in general studied dynamic
composition of complex services. Dynamic
composition is needed in order to dynamically
provide new services in an intuitive form based on
the semantics of the service. This kind of
composition will be vital especially in flexible
design of automation and control systems.
According to Fitzgerald et al. (2006) and
Kontogiannis et al. (2007) this is also of importance
in achieving third generation service-oriented
systems supporting context-determined composition
in a dynamic ad-hoc manner.
To be able to integrate design processes and
information in general both semantics and mapping
rules are needed to allow utilization and combination
of information sources. The Semantic Web is about
data and its meaning (W3C). Firstly it is about
interoperability in integrating and combining
information from diverse sources. Secondly it is
about relations between data and real world objects.
This allows, for example, information in different
formats to be interpreted concurrently at run-time.
The research and efforts contributed to the semantic
web are needed also to support the concepts of cloud
computing and operation models such as software as
a service (SaaS) in heavily integrated application
environments. In this perspective, both the solutions
ICEIS 2009 - International Conference on Enterprise Information Systems
220

provided as services and the data information itself
require semantics for integration and utilization.
3 CURRENT PRACTICES AND
CHALLENGES
Automation and control systems, manufacturing
execution systems (MES) and related information
systems are vital for smooth production in industrial
manufacturing. Requirements posed on these
systems are increasing as more functionality is
needed and as system integration and information
utilization extends all the way from the enterprise
level to field devices. The rapid development of
emerging technologies and new standards requires
designers to stay updated and to constantly develop
their professional skills. Design time work support
for design decisions generally relies on previous
experience and the support from colleagues within
the same team.
All this puts heavy pressure on engineering work
developing automation and control systems. The
applications easily become large and complex,
consisting of subsystems and integrations to and
from various existing systems. As a result also
project management becomes challenging and
fulfilling all functional and non-functional
requirements demands careful design and
implementation. This is an issue also because parts
of the development may be outsourced. At the same
time quality requirements on reliability, correctness,
accuracy and maintainability, to name a few, are
increasing.
Figure 1: Interweaved design activities and information
exchange between project parties.
Figure 1 illustrates a collaboration scenario and
the transfer of necessary information between the
different parties in the development process.
Developing automation and control systems is, in
general, characterized by requiring years of
experience and professional competence. Also vast
amounts of design information need to be transferred
between development phases, e.g. application design
requires source information from process and
preliminary design phases. A lot of information is
available both during design and plant operation but
the challenge is to know what is needed and how to
use it. During recent years, XML has made its
breakthrough as the data transmission standard
partly because of its interoperability. Despite its
platform independent nature, it is not enough when
the amount of information increases and
corresponding data models with different structures
and semantics are used concurrently. This does not
only concern design time data, but also information
needed while the system is operating, for example
information on system components for maintenance
or modernization purposes. While information
systems evolve during the plant life cycle both data
models and the way of accessing information may
change in a manner that requires adaptation.
Profound semantics are therefore needed to provide
the meaning of the data to fully utilize the
information available in a reasonable manner.
The importance of validation and testing cannot
be stressed enough for mission-critical systems.
Increased functionality results in more complex
systems and the validation and testing of design
decisions and implementation becomes challenging.
In cases where simulators and testing tools are
utilized as a part of the development the solutions
are mostly based on custom applications for specific
cases. The use of specialist services as a part of the
development is therefore laborious in most cases
while there is not any easy way to integrate
outsourced expertise into the development process.
In the automation and control engineering
industry it is common to use subcontractors in larger
projects. There are also firms offering specialized
services, such as simulation or knowledge on
advanced control methods. However, these are
typically limited to specific parts of the system,
independent subsystems or certain phases in
development. For example, the initial design phase
is performed by a subcontractor and only the final
design artefacts or the outcome is returned and
integrated manually or semi-automatically to the
following development phases. The same applies to
integrating design artefacts such as subsystems or
application components from different parties. Also
the simulation of these subsystems or other critical
elements of the application requires a lot of
additional work. For instance, even a simple
verification of impacts caused by a change in design
can be demanding. This shows that work processes
are not accommodated with current design systems
so they could be managed autonomously and so that
A SERVICE ORIENTED ENGINEERING APPROACH TO ENHANCE THE DEVELOPMENT OF AUTOMATION
AND CONTROL SYSTEMS
221

external services could be utilized effectively as a
part of the development process.
Manufacturing systems and production plants
generally have a long lifespan that requires various
supporting activities for optimal operation.
Automation and control systems are usually
connected with a large number of existing systems
such as ERP and systems related to manufacturing
and quality control. Examples of associated systems
are also condition monitoring and maintenance
systems. For instance, if maintenance is outsourced
to a third party the service provider is typically only
informed of events reported by monitoring systems
at the plant. The service provider may have to use
many existing systems at the plant to access all
necessary information, including design data, in
parallel with the service providers own systems.
Although expensive integrations are made between
systems, well-defined information management is
hard to achieve and the effective use and integration
of services is difficult.
4 ENHANCED DEVELOPMENT
PROCESS
To meet the challenges presented some
improvements could be made. Most of the changes,
however, can be limited to design systems and IT
systems involved so that actual work processes
merely have to be defined more precisely. The aim
is, although, to enable a more service based
operations model during system development that
also supports further life cycle activities.
4.1 A Service based Approach to
Improve Information Exchange
In distributed engineering environments crossing
organisational boundaries design information and
development artefacts are fragmented into different
systems and databases. The ideal solution would be
to use uniform standards and information models for
instant interoperability and easy integration. This
will presumably never be a feasible solution as there
are different needs depending on the system under
design and the task being performed. Therefore the
integration of design data should be made on a
deeper semantic level so that this information could
easily be combined and mapped to suitable
information models for the parties concerned. This
would also benefit all other supporting activities
such as simulation, testing and special expert
services.
Open and compatible standards are a basis for
interoperability between new systems, but the large
number of systems used also contains old legacy.
Therefore customized adapters and service wrappers
are needed to support existing systems and data
models. The technology exists to integrate systems
and to bring more semantics to service discovery,
information usage and the handling of work
processes. A service oriented architecture utilizing
semantic web technologies would provide means to
make the design of automation and control systems
more flexible and effective. It would also enable the
use of commercial services as part of the application
development process. A shared service framework
with standards based communication procedures
would be of common interest to support the
integration of design tasks and transformation of
design information from one system to another. This
would enable engineering companies to specialize
and offer their expert services for others to utilize as
a part of the development process.
To be useful the service framework should be
easily expandable with new services and allow
mappings to new data sources and information
models. For instance, if design information was
semantically annotated, vendors of devices and
components could offer related information and
special knowledge automatically to support the
development process. Empiric knowledge from
previous experiences could also be utilized this way
or it could be offered to partners and sub-contractors
as a service that automatically provides supporting
information during development.
Mapping and integration
Service
Provider A
Preliminary Design
(concrete service)
Simulation
Services
Design data
Service
Registry
Engineering and
workflow management
Process Design
(concrete service)
Application Design
(concrete service)
Service Provider B
Figure 2: Service based design and information
integration.
Figure 2 illustrates a service-oriented
development architecture with a subset of
engineering activities distributed into service units
representing specific development phases. The
figure also presents information exchange enabled
through mappings on a common semantic level
ICEIS 2009 - International Conference on Enterprise Information Systems
222

between services as well as integration of simulation
services to the design data.
4.2 Integration of Expert Services to
the Development
There are a wide variety of services that could be
enabled through this kind of a framework. Testing
design and implementation, for example, is
troublesome and it is important to find errors as
early in the development as possible. Simple
syntactic testing is possible but more advanced
testing and verification could be enabled through
intelligent services. These services can be automatic
and utilize a wide range of background information
making deductions. For example, statistics gathered
from similar systems already in production,
information on previous similar designs or
specifications on devices and components can be
used to evaluate design. The services could also be
performed manually by a third party specialist, e.g.
an experienced engineer. However, on a service
description and interface level it should integrate
into the development process as any other
engineering service.
Simulation is also an important activity during
development, especially for critical systems.
However, simulations are not used as widely as
possible partly because the creation of simulation
models is labour-consuming and requires special
expertise. If design information were available in a
format generally understandable, the automatic or
user-assisted generation of simulation models would
also be possible. This would enable the use of
simulation services to enhance the quality of the
system under design. Design models representing
the control application behaviour, for example,
would be useful to be simulated.
The use of design time information is worthwhile
also in supporting life-cycle services such as
maintenance. The share of repairing maintenance
compared to preventive or predictive maintenance is
often too high in the manufacturing industry. Design
information on system structure and components
used is valuable evaluating and preparing
maintenance schedules. Combined with advanced
condition monitoring systems problematic parts of
the systems can be analyzed and identified. In
addition, the information found may also be of
significance for future systems to be designed.
4.3 Improved Design Workflow
Management
A service-oriented approach organizing service
activities provides better management of involved
activities and improves much needed process agility.
Advanced business process modelling techniques
and modelling languages provide tools for analyzing
and orchestrating processes composed of services.
High-level rules describing workflows and the use of
advanced reasoning for automatic service discovery
and utilization bring flexibility and efficiency to the
development process. For example, invoking testing
or simulation services after changes to the design
could be automatic, either for the whole systems or
only a part of it.
Figure 3: A subset of engineering activities, corresponding
service definitions and service compositions.
Figure 3 presents some engineering activities and
their relationship to services in a development
workflow. To be able to implement tasks in the
development process presented the offered abstract
higher level services need to be defined with clear
and concise service level agreements for discovery,
binding and invocation. These abstract services may
then be composed of smaller services such as Web
services or activities performed by humans. The
challenge is to provide well defined service
interfaces, to bind them to services on different
levels of granularity and to control the activities in
the development process.
The key benefits of organizing design activities
into services and a common framework are in
interoperability, work process management and new
business opportunities for utilization of specialized
expert services. Based on a common service
framework, services could be customized for each
company, based on individual needs with common
building blocks without huge development
resources. This would allow businesses to specialize
in special sectors and make possible the utilization
of this special expertise as a part of the development
with feasible resources and efforts. This would also
imply the transfer of business models and invoicing
practices from licenses and agreements to be based
A SERVICE ORIENTED ENGINEERING APPROACH TO ENHANCE THE DEVELOPMENT OF AUTOMATION
AND CONTROL SYSTEMS
223

on service usage. A service-oriented approach based
on a common, sound information models also
enables separation of concerns between different
parties and engineering disciplines.
5 CONCLUSIONS AND FUTURE
WORK
System integration and the broader utilization of
information is becoming the cornerstone of effective
operation in all businesses. This paper has presented
some concepts for a new operations model
implementable as services with semantics to provide
enhanced information exchange between
engineering enterprises and support during the
design of automation and control systems.
A service based operations model can offer a
new level of flexibility and interoperability, and
support agile business practices and efficient
management of design work processes. A
development process based on services will also
enable new business opportunities for specialized
expert services and allow integration of these
services to the engineering process easily. The
automation and control domain has also established
practices with suitable standards and information
models for the development process to be
implemented in a service-oriented fashion. A
service-oriented approach based on a common
information integration layer will also support
various other design approaches, such as model-
driven methods for example, to be utilized in the
design process.
In order to fully utilize a service framework in
the development process many obstacles still have to
be overcome. Mapping services to business
processes needs further research concerning
infrastructure and strategies aligning processes with
service components. In addition service definitions
and service level agreements need further
investigation and considerable domain specific
input. The run-time discovery and composition of
Semantic Web services also need established
practices and there are many ongoing efforts
addressing this, e.g. OWL Web Ontology Language
for Services (OWL-S). The adoption of service-
oriented systems depends among other things on the
usability and the added value the associated process
models bring. Special attention is needed to support
a gradual transition from current practices to a
service based operations model.
The information content in design systems, for
example, is diverse and the amount of data is huge.
Efforts are needed to be able to utilize the
information, e.g. by semantically describing the
data. Modelling of complex design processes is also
needed in order to manage distributed design
activities in a more automatic manner. Knowledge
modelling and information management is needed to
be able to control information flows in new flexible
design environments. Therefore proof of concept
solutions with implemented services and process
management are needed to evaluate the functionality
and benefits of the service based operations model.
This paper has presented some concepts and
observations that are being explored in ongoing and
commencing research projects.
REFERENCES
Bergman, R., Borden, C.S., Zendejas, S., 2002. Automated
workflow for engineering services, IEEE Aerospace
Conference Proceedings, 5, 2553-2567.
Fuji, K., Suda, T., 2005. Semantics-Based Dynamic
Service Composition. IEEE Journal on Selected Areas
In Communications. Vol. 23, No. 12, 2361-2372.
Fitzgerald, B., Olsson C.M. (eds.), 2006. The Software
and Services Challenge. Contribution to the
Preparation of the Technology Pillar on Software,
Grids, Security and Dependability in the European
Union 7
th
Framework Programme. 30 January 2006.
Jammes, F., Smit, H., 2005. Service-Oriented Paradigms
in Industrial Automation. IEEE Transactions on
Industrial Informatics, Vol. 1, No. 1, 62-70.
Kontogiannis, K., Lewis, G., A., Smith, D., B., Litoiu, M.,
Müller, H., Schuster, S., Stroulia, E., 2007. The
Landscape of Service-Oriented Systems: A Research
Perspective. International Workshop on Systems
Development in SOA Environments, 2007. SDSOA
’07: ICSE Workshops 2007.
Lee, J.-K., Park S.-W., Kim H., S., 2007. Development of
Engineering Framework and Portal for the Integration
of Cooperative Engineering Processes. Proceedings of
the 2007 11
th
International Conference on Computer
Supported Cooperative Work in Design, 435-439.
Lee, J.-K., Kim H., S., Park S.-W., 2008. Extended BPEL
System for e-Engineering Framework Considering the
Characteristics of Mechanical Engineering Services.
IEEE International Conference on Services
Computing, 2, 481-484.
Thramboulidis, K., C., Doukas, G., Koumoutsos, G., 2008
A SOA-Based Embedded Systems Development
Environment for Industrial Automation. EURASIP
Journal on Embedded Systems, Vol. 2008, No. 1.
W3C Semantic Web Activity, W3C. Retrieved August 7,
2008, from: http://www.w3.org/2001/sw/.
ICEIS 2009 - International Conference on Enterprise Information Systems
224