WEB INFORMATION SYSTEM: A FOUR LEVEL
ARCHITECTURE
Roberto Paiano, Anna Lisa Guido, Leonardo Mangia
Dipartimento di Ingegneria dell’Innovazione-University of Lecce, Arnesano Street 73100 Lecce, Italy
Keywords: Business Process Management, Information System architecture, Web Information Systems Methodology.
Abstract: Business processes are playing a very
important role in companies and the explicit introduction of them in
Information System architecture is a must. According to the interest shown towards Web Application it is
important to introduce a new web-oriented class of software, which is able to gives to the manager the
possibility to operating directly with the process (we will talk about process oriented WIS - Web
Information System). It is necessary to replace the three-level logic of the traditional application
development (Data, Business Logic, Presentation), that hides processes in the Business Logic, with a four-
level logic that allows to separate the process level from the application level: definition and management of
the processes will be not tied solely to business logic. Our research (work in progress) focus is on an
innovative framework (software architecture and methodology) for Information System development that
links together the know-how acquired in Web Application design and in the process definition concepts.
1 INTRODUCTION
Industries have made great efforts in the last years to
improve their Information System which today is not
only a tool for activities automation but also a
powerful tool that enable managers to analyze their
company’s activities and to take timely decisions in
order to correct possible management errors. To
reach this goal, companies abandoned the vertical
vision that locates business logic in functional areas
and orients them toward a transversal vision that,
following process logic, tries to improve the
management exploiting as much as possible the
existing resources (software, hardware, actors and so
on). Process vision is a good way to guarantee to the
manager a full business activity control and to
increase flexibility degree in Information System
management. The attention paid to processes brings
to the necessity to integrate the existing applications
inside the company with the goal to meet managers’
goals. Although process vision is a good idea, today,
processes are only in the mind of the business
analyst and are not explicitly taken into
consideration by IT experts: business experts define
the process flow but, when the information system is
implemented, processes are hard coded in the
application without well defined guidelines, so,
developers are free to take decisions in order to
reflect the process logic. These decisions often
distort the original process logic in the final web
application. In this way a change in the process flow
means a change in the application code, so it is hard
to obtain “on the fly” a change in the information
system in order to meet the flexibility that the
manager needs. To obtain a new vision of the
process it is very important to separate the process
logic from the business logic and to provide to the
manager an horizontal view of the processes that
take in consideration several people who belong to
different division of the company. It is important to
introduce a new class of software that allows the
manager to define, to implement and to execute
processes. In order to do this it is essential to have a
tight cooperation between IT and Business sectors of
the company so that business requirement can have
in a brief time answers from the IT sector. Another
important consideration is that an Information
System must be addressed towards the horizontal
management of the company and must be delivered
on the web (we talk about WIS-Web Information
System) in order to reach several users and to
integrate services of other companies. Our research
work, starting from a Web Application design
177
Paiano R., Lisa Guido A. and Mangia L. (2006).
WEB INFORMATION SYSTEM: A FOUR LEVEL ARCHITECTURE.
In Proceedings of the First International Conference on Software and Data Technologies, pages 177-182
DOI: 10.5220/0001321901770182
Copyright
c
 SciTePress

methodology and using methodologies of process
analysis, design and representation already existing,
propose to define a new methodology that by linking
the know-how of both, drives the designer towards
the design of a WIS. Our goal is to provide not only
a methodology but also a set of tools that allow to
pass from the representation of the process to the
final Web Application where process can be
executed.
In the next section we present the state of the art
of the methodology that try to address the problem
and of the tools actually on the market. In the section
3 we present the importance to introduce the fourth
level in the overall information system architecture
and in section 4 we explain the architecture
proposed and a methodology that try to address the
problem. We, then, apply the methodology to a case
study. In section 5 we present conclusion and future
works.
2 BACKGROUNDS
In order to design and to implement a Web
Information System (WIS) it is important to address
two different problems:
• The necessity to take under control the web
application design and therefore the necessity to
take in consideration, in the design phase,
information, navigation, presentation and
transactional aspects typical of Web Applications
(Baresi, Garzotto, Paolini, 2005)(Baresi,
Garzotto, Paolini, 2001).
• The necessity to represent and to manage
business activities according to process logic.
Currently Web Information System design
separates the problems related to Web Application
from the problems related to the process
management. Current attempts to consider process
design in existing WA design methodology fail to
explicit separate process design from typical Web
Application design tasks. Moreover, processes
implementation is in the mind of the developers
which hides the process in the Web Application in
some way difficult to understand and to modify.
There are different approaches in the scientific
community that take into consideration the process
design explicitly in the Web Application design
methodology. An example is in (Koch, N., Kraus,
2003) where the authors attempt to extend web
application methodologies such as OO-
H(
Cachero,2003) (Gómez, Cachero, Pastor, 2001) and
UWE (Koch, Kraus, 2003) with process design. The
authors define a process model (made by UML
activity diagram) that allows to design process flow.
After the process model definition, the navigation
model, designed without take into consideration the
process, is integrated with process model in order to
allow the final user both to navigate in the general
information and to execute process steps. The
Navigation model defines a process entry and exit
points between tasks. In this interesting approach the
design start from the navigation model without
taking into consideration, in a separate way, the
information involved in the Web Application and in
the process too. Another approach is proposed in
WebML (Brambilla,2003); this approach extends its
data model to define the process but the approach is
limited to very simple process without a complex
logic and the authors say that this approach is not
valid when the logic complexity is high. As it
regards the technological point of view, currently
there are on the market a lot of frameworks that
allow to define processes and to develop
applications that implements them: these
frameworks are both complex and expensive. As an
example, Filenet P8 (http://www.filenet.com), W4,
iLog, PopkinSoftware and Fuego
(http://www.Fuego.com,http://www.ilog.com,http://
www.popkin.com, http://www.w4global.com,). The
main problem related to these frameworks is that
they are difficult to apply in a small-to-medium
sized companies (target of our research work)
because they are very complex and therefore they
require high investments both to purchase hardware
and to training people in the company that will adopt
these frameworks. These frameworks, also, does not
consider any design methodology or guidelines in
order to design the final web application.
3 THREE LEVELS VS. FOUR
LEVELS APPLICATION
ARCHITECTURE
The introduction of a new class of software that
allows to take under control in an explicit way the
definition and management of processes imposes to
modify the current application architecture. Today,
applications are usually structured in a three level
architecture: there is a Data level made up of
centralized or distributed databases where
information are stored; Application level where
there is the company business logic that exploit and
manage data of the underlying level; finally
Presentation level that allow the end user to interact
with business logic. Since the ‘90s, process logic
ICSOFT 2006 - INTERNATIONAL CONFERENCE ON SOFTWARE AND DATA TECHNOLOGIES
178

began to affirm with the BPR theory (Business
Process Reengineering) (Hammer, 1990) which
introduced the process as the way to improve the
company’s management. Nevertheless analysis and
re-engineering of processes is a preliminary phase of
the Information System realization or change, the
processes are, today, not explicit in company’s
management and thus it is very difficult to monitor
and to manage them. We think that it is very helpful
to separate the application level from process level
and, to do so, we introduce the process level that
allow to define and implements process in explicit
way (eventually re-using existing business logic)
and to monitor them. Process level will be the
contact point between business experts and IT
experts each of them with a specific goal:
• Business experts will focus on the processes and
in the direct process management in order to
modify the company’s work: a change to the
process will directly reflect on the application that
implements it without a full re-design and re-
implementation; to change process means to
change the management of the business logic and
not the business logic. In this way the application
obtained will be more flexible and will be focus
on the process logic and not on the business logic.
• From the IT experts point of view the code
maintenance is simplified: a change in the
business logic means a change in the application
level code without any (or with few) changes in
the Presentation and Process level; a change in
the process will directly reflect a change in the
process level without modify the business logic
and with some changes in presentation level. In
the company, very often, there is the necessity to
change (or re-define) processes and not the
underlying business logic, for this reason the
introduction of a process level is very helpful. In
the process level processes may be defined; it is
possible to simulate its behavior and to change, if
needed, the design, The difference between
process level and application level is that in the
process level it is possible to define actors,
process flow, tasks, while in the application level
there is the implementation of each task (or
groups of task), the implementation of the access
right to each actor and so on.
Figure 1: The architecture of our research.
4 THE PROPOSED
INFORMATION SYSTEM
ARCHITECTURE
The architecture that we propose in our research
work is in Fig. 1. Our work is at an initial stage, so
we would like to explain here our idea without
provide too many technical details. First of all there
are three different parts: requirement analysis,
design methodology and the four-level application
architecture. After the requirements analysis, the
next step is to design the Web Application that
allows to take in consideration also the process
analysis. Our idea, explained in the next section, is
to define an innovative methodology that integrates
in an opportune way the IDM (Perrone, Bolchini,
Paolini, 2005) design methodology with the process
definition finding the contacts points between this
methodology and the process definition. Starting
from this new methodology will be possible, with an
effort that, we foresee, will be hard, to obtain a final
Web Application, automatically or manually
produced, that goes over the simple definition of
information contents and navigation between them
but includes also a process and the instruction of its
execution. In the process level it is possible to
represents the process (Process Definition) using
k
WEB INFORMATION SYSTEM: A FOUR LEVEL ARCHITECTURE
179

BPMN (Business Process Management Notation)
(OMG, 2006). The designer has ad-hoc tools, the
process editor (Guido, Paiano, 2006), that allow to
obtain as output the process model (Model) in a
formal language (OWL language (W3C, 2004)). The
editor, thanks to BPMN notation and to the efforts
made to provide a tool simply to use and to
understand both by business and by IT experts, is
independent from any analysis methodology we
think to use. After process definition it is possible to
validate the design (process validation) in order to
verify if the design follows the BPMN constraints
(as an example the process design infinite loop are
not accepted or each fork will be followed by a join
and so on). The output of the process editor will be a
formal representation of the process, the process
model, where it is possible to find all the
information about process. Process model will be the
input for the process engine tool (not yet
implemented) that manages the process execution
that is the process flow and the assignment of each
task to the right actor. The process engine will
interact with the application level and allow to
include in the process management also business
logic not yet implemented in the company. It is clear
that the methodology will help in the interaction
between different levels. In the architecture show we
refer also to the process simulation block (we think
to complete this tool within 6 months) that allows to
define Key Performance Indicator (KPI) and to
verify if the process meets the performance
requirements.
4.1 Web Information System Design
Methodology
The web application design methodology reflecting
the IDM methodology aspects is transversal to the
Data level, the Application level and the
Presentation Level. The methodology proposed links
both the know how obtained by the IDM design
methodology (so will solve the information,
navigation and transactional aspects typical of web
application) and the know how coming from the
process design knowledge. First of all we think that
two different experts will design, independently from
each other, the information system processes (using
BPMN notation) and the information and navigation
aspects (using IDM design methodology). In fact
process design will explain how companies work;
while information and navigation design will explain
what the information system provide to the final user
from the communicative point of view (in a few
word design the design of the user experience). In
our study we first evaluate the possibility to provide
a direct mapping between process task and IDM
design. We observe that it is impossible to provide a
general strategy to do so because the user experience
design is made up not only by information related to
the process (expressed as input and output of the
task) but it is often important to add other
information not directly connected to the process
Figure 2: BPMN design of the management of the flora-faun census process.
ICSOFT 2006 - INTERNATIONAL CONFERENCE ON SOFTWARE AND DATA TECHNOLOGIES
180

definition but strictly related to the user-experience
design. We think that, probably, it is possible to see
some contact points between information needed in
the process definition and information needed from
the user experience point of view. It is important to
cover also the navigational aspect typical of Web
Application taken in consideration in the IDM
methodology. When we consider processes, it is
important to think not only to the navigational link
related to the user experience (semantic
relationships) but also to the navigational link that
allow the final user to follow the process flow. In
our experience with several case studies, a mismatch
between process link and semantic link has been
covered by a process re-design and/or a user
experience redesign. Thus, this mismatch could be a
signal that there are some problems in the process
design and/or in the user experience design. These
aspects bring us to think that process links are a
subset of semantic links. The main difference
between process link and semantic link is that the
user is free to navigate between semantic links but
he/she is not free to navigate between process links
because there is an underlying process flow that
he/she must follow. Another important aspect is
related to the access structure: it is important to
define ad hoc access structures that, depending on
the state of the process, provide to the final user the
access to the process in the exact point where he/she
leaves from it in a previous session. Access structure
strictly related to the process flow would be added to
the access structure related to the user experience
design, so we may see in the methodology two types
of access structures: the access structure coming
from the user experience design and the access
structures coming from the process design.
4.1.1 A Case Study: Environmental
Monitoring
We apply our consideration about the four layer
application and about the methodology to a research
industrial project founded by the Italian Government
about environmental monitoring. In this project
business experts and IT experts teams design, in
separate way, both the user experience (through
IDM) and the business process (through BPMN) of
the overall information system. We present here, due
to space limitations, a fragment of the overall
Information System: the management flora-faun
census. The BPMN design of this process is in fig.
2. The sub-process “data collection” considers three
type of census but we are interested only to
“faunistic census”. The integrated design, made up
by the observation made in previous section, is
shown in fig. 3; it refers only to the coordinator user.
We observe that in the integrated design we can
recognize the four level architecture presented
above: the data level is the definition of the
information content and their relationship while the
Figure 3: The integrated design.
WEB INFORMATION SYSTEM: A FOUR LEVEL ARCHITECTURE
181

application level is represented with the property of
each operation act (we does not show these property
for the space limitation but in each operation
act(
) is well defined how it is implemented). We
can see the process level inside the integrated design
in the start process(
) in the intermediate event
(
) and in the process links (that links together
operation act with the name of the destination task).
Note that gateways are not presented in the
integrated design because they are defined in the
attributes of the operation act. It is very important to
observe that in Fig. 3 only the “protection circle” has
its own operation act but it is possible, in other case
study, to have operation act in several topic so the
process involve different topics but the mapping
between BPMN and IDM design is also valid.
Another observation is that there are several access
structure (near to the protection circle topic) in order
to take in consideration the entry point to the
application coming from the process design, these
are obviously not taken in consideration in the IDM
design before integration. It is important to highlight
that the process flow can be followed directly in the
integrated design, so a change in the process flow
will be immediately followed by a change in the
integrated design, and, thus to the final web
application that take in consideration both the
process and the user experience issues. This allows
to obtain the right flexibility required by modern
Information System.
Finally, the presentation level is not show in this
integrated design because we define this level
through another methodology (named E-WOOD)
more suitable for the presentation design.
5 CONCLUSIONS AND FUTURE
WORK
Modern Information Systems will be more flexible
and able to answer “on the fly” to the manager
innovation requirements. The only process analysis
and re-definition is not too much to meet the
Information System flexibility and, as consequence,
it is necessary to make explicit the process in the
WIS architecture. We add a fourth level in the
traditional three level Information System
architecture in order to explicitly manage processes.
We provide not only tools to manage process but
also a methodology that links together the know how
coming from process design and the know how
coming from the user experience design. A case
study is shown and we highlight where the four level
proposed are in the application design. Our research
work is not yet concluded, we work to refine the
methodology proposed and to complete the
development of tools in the process layer.
REFERENCES
Baresi, L., Garzotto, F, Paolini, P. (2001). Extending UML
for Modeling Web Applications. Proc. of 34th Annual
Hawaii International Conference on System Sciences
IEEE Computer Society. Island of Maui, Hawaii.
Baresi, L., Garzotto, F., Paolini, P. (2001). From Web
Sites to Web Applications: New Issues for Conceptual
Modeling. Proc. WWW Conceptual Modeling
Conference, Salt Lake City.
Brambilla, M. (2003). Extending hypertext conceptual
models with process-oriented primitives. International
Conference on Conceptual Modeling (ER), Chicago,
Illinois USA
Cachero, C. (2003). OO-H: Una extensión a los métodos
OO para el modelado y generación automática de
interfaces hipermediales. Retrieved October,
2005from
http://www.dlsi.ua.es/~ccachero/pTesis.htm
Gómez, J., Cachero, C., Pastor, O. (2001). On Conceptual
Modeling of Device Independent Web Applications:
Towards a Web Engineering Approach. IEEE
Multimedia Special Issue on Web Engineering. Vol.
8(2) pp. 20-32
Guido, A.L., Paiano, R. (2006). BPM:Overall Architecture
and Business Process Design Tool WebIst 2006
Setubal,Portugal ISBN978-972-8865-46-7
Hammer, M.(1990). Reengineering Work: Don’t
Automate, Obliterate. Harward Business Review. Vol.
68 pp. 104-112
Koch, N., Kraus, A. (2003). Modeling Web Business
Processes with OO-H and UWE. Third International
Workshop on Web-Oriented Software Technologies,
Oviedo, Asurias
Koch, N., Kraus, A.(2003). Towards a Common
Metamodel for the Development of Web Applications.
International Conference of Web Engineering, LNCS,
Springer Verlag Oviedo Spain 497-506
OMG, BPMN 1.0, (2006). Final Adopted Specification,
February 6, 2006 Retrieved April, 2006, from
www.bpmn.org
Perrone V., Bolchini D., Paolini P. (2005). A stakeholders
centered approach for conceptual modeling of
communication-intensive applications. Proc. of the
23rd annual international conference on Design of
communication: documenting & designing for
pervasive information pp: 25 – 33, ISBN:1-59593-
175-9
W3C, OWL Web Ontology Language Overview, (2003):
retrieved on http://www.w3.org/TR/2003/CR-owl-
features-20030818/
ICSOFT 2006 - INTERNATIONAL CONFERENCE ON SOFTWARE AND DATA TECHNOLOGIES
182