Oscar Mangisengi
BWIN Interactive Entertainment AG, Marxergasse 1B, A-1030 Vienna, Austria
SMS Data Systems, GmbH, Lilienbrunnen 18, A-1020 Vienna, Austria
Ngoc Thanh Huynh
Mayr-MeInhof Karton Gesellschaft m.b.H, Brahmsplatz 6, A-1041 Vienna, Austria
Keywords: Activity Warehouse, Business Activity Monitoring, Workflow Management Systems, On-line Analytical
Processing (OLAP), and Business Intelligence.
Abstract: An existing challenge is that organizations need to make business processes as the centrepiece of their
strategy to enable the processes performing at higher level and to efficiently improve these processes in the
global competition. Traditional Data Warehouse and OLAP tools, which have been used for data analysis in
the Business Intelligence (BI) systems, are inadequate for delivering information faster to make decisions
and to earlier identify failures of a business process. In this paper we propose a closed-loop BI framework
that can be used for monitoring and analyzing a business process of an organization, optimizing the business
process, and reporting cost based on activities. Business Activity Monitoring (BAM) as data resource of a
control system is as the heart of this framework. Furthermore, to support such a BI system, we integrate an
extracting, transforming, and loading tool that works based on rules and state of business process activities.
We show that some functionalities of our prototype are working well. The tool can automatically transfer
data into a data warehouse when conditions of rule and state have been satisfied.
In today’s fast-paced business environment and due
to regulatory compliance, organizations encounter
new challenges to increase an ability to make
adjustments quickly, to monitor and optimize
business activities and to early identify process
defects. The optimization of activities within a
business process aims at analyzing cost of activities,
so that an organization is able to improve
competitiveness, and to provide high quality services
to gain market. Furthermore, organizations need to
make business processes as the centrepiece of their
strategy to enable the process to perform a higher
level (Lopez-Claros 2006).
Information technology provides a highly
significant role for making better decision and
controlling business processes. A traditional
Business Intelligence system has been developed a
decade ago and is built on data warehouses (DWs)
as a data resource for On-Line Analytical Processing
(OLAP) (Codd et al., 1993) tools to make decisions
for high-level management. DWs store historical
data and data in a DW is organized as
multidimensional data (Kimball et al., 2002, Inmon
2002). DWs and OLAP tools offer strong data
analysis functionalities based on multidimensional
data and its hierarchies, such as rolling-up, drilling-
down, or drill-through. The concept of data
warehouse concentrates on data, however, in fact,
business processes of an enterprise or an
organization are processed from department to
department and run as a long–running transaction.
Such a process can be divided into discrete
activities. Activities, checkpoints, and events occur
within the process. As the result, granulated data is
created inside the process. Unfortunately, the
granulated data is not preferred to store in a data
warehouse. DWs are modelled to store transaction
end counts, and are not designed to capture events of
a business process. DWs store end counts, rather
than process checkpoints (Creese 2005). Therefore,
Mangisengi O. and Thanh Huynh N. (2008).
In Proceedings of the Tenth International Conference on Enterprise Information Systems - DISI, pages 210-217
DOI: 10.5220/0001693602100217
data stored in DW lacks process contexts and does
not represent events of the business process.
Furthermore, data warehouse as a part of BI system
is insufficient data resource to deal with the
emerging new challenges of business environment
In this paper, we identify two challenges for
rebuilding BI applications in our work. They are
given as follows:
A further development of BI applications
attempts to integrate workflow technology
(Sheth et al., 1999) and Business Activity
Monitoring (BAM) (Dresner 2002) into data
warehousing. A workflow of a transaction of a
business process is a long-running transaction
and in addition activities of the workflow have
to be monitored in detail. Therefore, a data
repository that is used to store event, activity, or
process data or event-oriented data within the
transaction of the business process is necessary,
and such a repository can be used as data source
for monitoring and detecting the events,
activities, and processes of the workflow. In
(Mangisengi et al., 2007) an activity warehouse
has been successfully implemented and the
event-oriented data stored in activity warehouse
can be used for efficiently monitoring business
processes in real-time and providing a better
real-time visibility of the business process.
The needs of establishing BI framework that
can integrate data resources and new
applications into a BI system are essential to
address business requirements and to improve
business processes, e.g., a solution must be
responsive to the demand of issue. Shorter
decision cycles require more flexibility.
Decision-making flexibility requires the ability
to perform continuous ‘what-if’ analytics and
ongoing scenario planning. It is inadequate to
deliver information faster to decision makers in
the form of reports, dashboards, or alerts only.
Moreover, it is essential for supporting for
evaluating decision alternatives. Building a
sense-and-respond capability in a process
context is critical to realizing continued
improvement in business performance (Morris
et al., 2007).
This paper proposes a closed-loop Business
Intelligence framework addressing the emerging
challenges to provide adaptive information and to
enhance decision-making cycles. The notions of a
closed-loop BI framework are able to early identify
and detect failures of activities in a business process,
to provide a closed-loop decision based on tactical
and control data and strategic data on the operational
data, to integrate data centric and process centric
into a unified system in order to work together to
provide an adaptive information, and to report
activities based costing.
In our approach we divide data resources for
establishing a closed-loop BI framework into three
different classes of data (i.e., operational, tactical
and control, and strategic data). Using a rules and
state-based ETL tool, they work together to provide
adaptive information.
This work is organized as follows. Section 2
outlines related work and contribution. Furthermore,
we present our motivation in Section 3. Then,
section 4 presents briefly an overview to derive an
activity warehouse based on the BAM requirements.
A closed-loop BI framework is presented in Section
5. Finally, conclusion and further work are presented
in Section 6.
Recently there exist research works in the literature
for the architecture of Business Activity Monitoring,
workflow management systems, and real time data
warehousing. The architecture of BAM is initialized
and introduced in (Dresner (2002), Nesamoney
(2004), Hellinger & Fingerhut (2002), White (2003),
McCoy 2001). The concept of process warehouse
has been introduced for different purposes, such as
in (Nishiyama 1999) a process warehouse focuses on
a general information source for software process
improvement, (Tjoa et al., 2003) introduces a data
warehouse approach for business process
management, called a process warehouse, and in
(Pankratius & Stucky 2005) Pankratius and Stucky
introduce a process warehouse repository.
Furthermore, in relation to data warehousing,
(Schiefer et al., 2003) proposes architecture allows
transforming and integrating workflow events with
minimal latency providing the data context against
which the event data is used or analyzed. An
extraction, transformation, and loading (ETL) tool is
used for storing workflow events stream in Process
Data Store (PDS).
Based on the successfully implementing Business
Activity Monitoring for integrating enterprise
applications (Mangisengi et al., 2007), in this paper
we continue our work to face the second challenge
to reach a future, powerful business intelligence
We argue that a traditional BI system, which
includes DWs, OLAP applications and some Data
Mining/Reporting tools, is insufficient to meet the
emerging challenges, e.g. to increase an ability to
make adjustments quickly, to monitor and optimize
business activities, and to identify process defects
earlier. Although in the last few years there are
many powerful tools developed for specific tasks in
these areas, however, there is still need a
controllable and monitorable BI framework. It
cannot only adapt to the now-a-day quickly changes
in business requirements and processes, but also
provide a way to assess and optimize business
processes. Within our BI framework we integrate the
traditional BI system with BAM and other
technologies (i.e., workflows and business activity
monitoring). Furthermore, the adoption of new
technologies for BI system changes the way of
business management, assessment, and optimization,
e.g., shorter life cycle of a business assessment,
accelerating the information flows within the
system. The intention of such a change is on the one
hand to provide better decisions at the right time
(i.e., strategic and tactical decision in the
operational), on the other hand to measure
performance and optimize business processes so that
the cost of business processes can be analyzed (i.e.,
activity based-costing).
On the whole, a BAM system is the heart of the
control system. In this section we briefly present an
overview of a BAM system. The detail of this
system is given in (Mangisengi et al., 2007). To
monitor and control business activities, a BAM
system must be able to capture events and activities
of a business process. Additionally, this paper also
address a system, called activity warehouse which
depends on various requirements, discreteness of a
business process, business process management,
workflow, and the repository of BAM. We
respectively present the requirements in the
following sub sections
4.1 Discreteness of a Business Process
In order to monitor and control activities within a
business process, we divide a business process into
discrete processes. The discrete processes are
intended to obtain granulated data that represents
activities of a business process, so that a business
process can be analyzed and reported in very detail.
To discrete a business process, we need a conceptual
hierarchical structure of a business process given in
Figure 1 (Mangisengi et al., 2007).
Figure 1: The conceptual hierarchy structure of a business
Figure 1 shows a business process with a set of
activities represented from the highest level to the
lowest one. In this example the activity 1.3.2 cannot
go further to the next activity 1.3.3 activity if a
failure exists, it then goes back to the previous
activity 1.3.1. The conceptual hierarchy structure of
a business process given in Figure 1 can be
represented as follows:
A business process can be organized into a
hierarchical structure that represents different
level of importance from the highest level to the
lowest one, or vice- versa.
A business process may be decomposed into a
set of processes. A process may consist of a set
of sub-processes, and a sub-process includes of
a set of activities.
An activity is the lowest level process of a unit
transaction. It can be represented as three-
dimensional workflow given in Section 4.3.
4.2 Business Process Management
BPM aims at enhancing the business efficiency and
responsiveness of an organization and, at optimizing
the business process of the organization. BPM has
closed relationship to the business strategy of an
organization. BPM requires essential data and its
intention for enhancement and optimization as
Strategic data is to provide the result of an
organization that can be achieved and its
ICEIS 2008 - International Conference on Enterprise Information Systems
hypotheses. Also, it can be supported by the
Tactical data aims at controlling and monitoring
business process activities and its progress in
detail and supports a contextual data.
Business metrics data is to support the strategic
improvements for the higher level goals. In
addition, it supports departments and teams to
define what activities must be performed.
4.3 Workflow Management
To derive the repository of a BAM system, called
activity warehouse, we approach a process and state
workflow management. Depending on the business
requirements, a specific workflow will be used for
managing a business process, however, in general
there exist two characteristics of workflow that are
included in the activity warehouse to store data in
the particular context of business process activities
as follows:
Tracking Activity. The tracking activity deals
with the checkpoints of business process
activities of a unit transaction. It provides the
history of activities of a unit transaction.
Status Activity. The status activity represents the
status of a unit transaction after the execution of
a business process activity. The current status is
also used by an actor to address and execute the
next activity of the business process and in
addition to arrange the executions of workflow
in its correct order.
In our point of view an activity of a business
process can be represented as three-dimensional
workflow as follows:
Action. An action is represented by the method
of a particular activity and is corresponded with
an actor. Activities may be assigned to actors,
applications, or system queues based on rules.
Process. A process is a network of activities,
with rules to control the start and exit conditions
for each activity and the data flow between the
activities. It defines the business process
activities and the sequence in which they are to
be performed.
Actor. An actor is defined as a person or an
intelligent agent who or which respectively
executes a particular action. Furthermore, the
actor has a role and an organization.
Other requirements used for optimizing business
process are measurement data. The measurement
data consists of two levels of data, namely macro
and micro. In the macro level data represents end
count of a unit transaction within a business process,
whereas in the micro level data represents activities
within a business process. Analysing/mining the data
in the micro level we can detect failures in a
business process and/or to provide different tactical
decisions. It supports time and cost efficiencies. The
time efficiencies aim at optimizing business
processes and consist of cycle time, work time, idle
time, transit time, queue time, and set up time,
whereas the cost efficiencies are intended for
calculating costs based on these activities in relation
to the business process.
4.4 BAM Repository
A BAM repository (activity warehouse) is given in
Figure 2. An activity warehouse provides data
resources that are rich in process context and event-
oriented data. Figure
2 shows the relationship
between the OLTP system (i.e., the unit transaction
table) and the activity warehouse table.
The model consists of the Activity Warehouse
table and a set of dimension tables (i.e., Actor,
Activity, and Status). The activity warehouse table
consists of unit transaction identity, a set of
dimension tables, and a set of attributes for
measurement and optimization purposes, such as
cost and time efficiencies. The table activity
warehouse is represented as follows:
Status_ID, Activity_ID,
Actor_ID, StartOfAction,
EndOfAction, Duration, EntryDate,
CycleTime, WorkTime, IdleTime,
TransitTime, QueueTime,
SetupTime, Cost)
A set of dimension table consists of the Activity,
Actor, and Status dimensions.
Actor(Actor_ID, FirstName, LastName,
Activity(Activity_ID, Description,
SubProcess, Process)
Status(Status_ID, Description,
The activity dimension categorizes a business
process until the lowest level. The Process attribute
represents the highest level of the business process
and the Activity_ID and its description attributes
represent the lowest level of the business process.
Figure 2: Activity warehouse and unit transaction.
This section focuses on discussing a closed- loop
Business Intelligence framework in detail and how
its components can interact to each other’s so that
they can provide adaptive information. In relation to
our previous work, the BAM system can be used for
efficiently monitoring business processes in real-
time and provide a better real-time visibility of the
business process. We integrate a BAM system into a
traditional BI system, combine all traditional BI
component and BAM into a new framework that has
a robust controlling characteristic and a closed-loop
BI framework is obtained. The notion of the new
framework enables enhancing decisions, such as:
Monitoring, identifying and detecting defects of
activities earlier in a business process.
Providing decisions created from control and
tactical, strategic, and operational data that are
given by a closed-loop system.
Integrating data centric and process centric into
a unified system in order to work together to
provide an adaptive information.
Reporting activities based costing.
This framework consists of three main
components, named operational, BAM, and data
warehouse subsystems. They provide three different
data resources, namely operational data, control and
tactical data, and strategic data respectively. All data
of the system represents data centric and process-
oriented data. With the supporting of BAM, the
event-oriented data now will be used for controlling
the ETL tool to extract, transform, and load data
from operational system to DW. On a whole a
closed-loop BI framework is shown in Figure 3.
Figure 3: A closed-loop BI framework.
Figure 3 shows that a closed-loop BI framework
consists of an application layer, strategy decision
layer, and operational and tactical decision layer and
its workflow. The application layer consists of
different applications for different management
levels, such as activity monitoring, OLAP tools, and
OLTP applications. The strategic decision layer
comprises data warehouse and an ETL tool, whereas
the operational and tactical decision layer includes
OLTP system data and BAM system.
In the following sub-sections we demonstrate to
addressing the notions.
5.1 System Interaction and Data Flow
Dependent on information or decisions that will be
generated, system interaction and data flow can be
categorized into three classes:
Strategic Decisions. To generate strategic
decisions, operational system interacts with data
warehouse, where the interaction moves data
from operational system to data warehouse
using an ETL tool. This mechanism is similar to
in a traditional data warehousing. However, the
ICEIS 2008 - International Conference on Enterprise Information Systems
ETL tool of this framework uses different
approach that is given in Section 5.3.
Tactical Decisions. Tactical decisions are
generated by monitoring applications, where
data is taken from the activity warehouse. These
decisions monitor failures of business processes.
Monitoring applications interact with BAM
Operational Decisions. To generate decisions,
the framework apply two approaches, namely
from operational applications and from closed-
loop approach. The second uses tactical
decision and strategic decision for improving
processes in the operational application.
A system interaction and data flow in our closed-
loop BI framework is shown in Figure 4.
Figure 4: System interaction and data flow in a closed-
loop BI framework.
To support the system interaction and data flow
in the framework, we have implemented a prototype
that is divided into some modules and its basic
1. Business Process Module. It contains a business
process scenario namely:
o What-if
2. Monitoring and BAM Module. This module
consists of the following functionalities:
o Refreshing
o Browsing
o Throwing events
o Auto-detecting
3. Rule-based ETL Module. It consists of
functionalities as follows:
o Trigger
o Procedures
5.2 Business Process Improvement
To quickly and correctly improve or optimize a
business process organization needs not only
flowchart tools or methodologies, but also a costly
time and men-power analysis process. The both
aspects are available in this closed-loop BI
framework. Analysing data of all states and
activities of a business process organization can
identify the weak points or undesired properties of
this process. Additionally this framework supports
on-demand activities analysis, which allows
organization making a correct decision in a short
period of time.
5.3 Rules and State-based Extraction,
Transformation, and Loading
Rules and state-based extraction, transformation, and
loading tool are an important approach of this
system. The ETL tool works dependent on rules and
states of activities within a business process. Thus, a
regular time is not required to move data from
operational system to data warehouse. The ETL
automatically works when a specific state of an
activity within a business process is achieved. The
characteristics of the ETL tool are flexible and more
5.4 An Extension of On-Line Analytical
Traditionally OLAP tools aggregate data in DW for
rolling-up or drilling-down functionalities. To
monitor activities and detect failures within a
business process, we extend OLAP functionalities,
so that they can be used for accessing data micro
level in an activity warehouse and for summarizing
and calculating costs and performances based on
activities for a unit transaction. Additionally, new
functionalities to calculate and to report costs- and
performances-based on activities or processes are
needed. To navigate data stored in activity
warehouse, we provide metadata given in Figure 5.
It applies the multidimensional graphical notation as
it is developed by (Bulos 1996) using the ADAPT
modelling tool developed by (Totok & Jaworski
Figure 5 shows that each dimension contains the
All hierarchy level. For example, to obtain micro
level data based on dimensions, such as process,
state, and activity, we navigate the hierarchy level of
the Process dimension All Æ Process Æ Sub-
process Æ Activity, the State dimension (i.e., All Æ
Category Æ State) and the Actor dimension (i.e., All
Æ Role Æ Actor) respectively. On the contrary, to
Business Process
obtain macro level data, we navigate the hierarchies
on the opposite direction. In addition, using this
metadata navigation, On-Line Activity-based Costing
and Performance can be monitored and summarized
in real-time for one or more unit transactions.
New functionalities of OLAP applications have a
strong relation to different level managements. For
example, applications, such as activity monitoring,
costing, and performance, can be categorized into
some high-level management applications. They are
executed by different decision-makers dependent on
their roles and securities, however, in this paper we
do not discuss on the security issues in any detail.
Figure 5: An extended of OLAP for navigating micro level
5.5 Comparison between a Traditional
and a Closed-Loop Business
Intelligences Frameworks
We summarize a comparison between the traditional
BI and a closed-loop BI framework. The comparison
is given in
Table 1.
Table 1 shows that the closed-loop BI framework
provides some extending functionalities as well as
provides a better environment for decision-making
compared to the traditional BI framework. An
important feature of the closed-loop BI is that an
ETL tool works based on rules and states of a
business process.
Table 1: A comparison between traditional Data
Warehousing and the proposed a closed-loop OLAP
Feature Traditional BI A closed-loop BI
Operational and
strategic data
Operational, strategic,
and tactical and
control data
ETL Working based
on regular basis
Working based on
rules and states
OLAP Summarization,
monitoring, failure
detection, cost
Aggregated data,
macro level data
Aggregated data
macro, and micro
level data
Decision Strategic
Tactical and strategic
The aim
of data
Business Business,
In this paper we have proposed a closed-loop On-
Line Analytical Processing system to bring adaptive
information to a Business Intelligence framework.
Our approach is based on three data resources (i.e.,
operational, tactical and control, and strategic) to
provide the framework given by operational system,
Business Activity Monitoring (BAM) system, and
Data Warehouse respectively.
The advantages of the closed-loop OLAP BI
framework are to earlier enable detecting failures
within a business process, to monitor and optimize
the business process in detail, and to improve a
quality of the business process.
Integrating the closed-loop framework into
current traditional BI framework in real-world
business applications becomes a challenge for
enterprises and organizations because of a different
life cycle of the development process. Based on our
experiences, since BAM system is the centrepiece of
the closed-loop BI framework, first a robust BAM
requirement analysis and business processes of an
organization are completely necessary and second
BAM system must be well coupled to an operational
Our implementation is based on the Java
technology and it’s Service-Oriented Architecture.
Currently we are integrating BAM system and DW
into unified information to bring sense-and-response
ICEIS 2008 - International Conference on Enterprise Information Systems
information as well as rules and state- based
extraction, transformation, and loading system and
incrementally implementing the framework.
We believe that a closed-loop Business
Intelligence framework plays a major role in the area
of enterprise application integration in the near
future. Furthermore, we continuously investigate on
security issues of the new framework in real-world
Bulos, D., 1996. A New Dimension, in Database
Programming & Design, 1996.
Codd, E.F., Codd, S.B., & Salley, C.T., 1993. Providing
OLAP (On-Line Analytical Processing) To User-
Analysts: An IT Mandate, E.F. Codd & Associates,
White Paper.
Creese, G., 2005. Volume Analytics: Get Ready for the
Process Warehouse. DMReview, April 15, 2005.
Dresner, H., 2002. Business Activity Monitoring: New Age
BI?, Gartner Research LE-15-8377, April.
Hellinger, M, & Fingerhut, S., 2002. Business Activitity
Monitoring: EAI Meets Data Warehousing. EAI
Inmon, W., 2002. Building the Data Warehouse, John
Wiley & Sons.
Kimball, R., Ross, M., & Merz. R., 2002. The Data
Warehouse Toolkit: The Complete Guide to
Dimensional Modeling, John Wiley & Sons.
Lopez-Claros, A., 2006. The Global Competitiveness
Report 2006-2007, The Global Competitiveness
Index: Identifying Key Elements of Sustainable
Growth, 2006.
Mangisengi, O., Pichler, M., Auer, D., Draheim, D., &
Rumetshofer, H., 2007. Activity Warehouse: Data
Management for Business Activity Monitoring.
Proceedings of the 9
. International Conference on
Enterprise Information Systems, Funchal-Madeira,
McCoy, D., 2001. Business Activity Monitoring: The
Promise and the Reality, Gartner Group, July.
Morris, Vassett & Fleming, 2007. Improving Business
Performance through Process Visibility. IDC Analyze
the Future, May 2007.
Nesamoney, D., 2004. BAM: Event-Driven Business
Intelligence for the Real-Time Enterprise. DM
Review, Vol. 14, Number 3.
Nishiyama, T., 1999. Using a Process Warehouse Concept
A Practical Method for Successful Technology
Transfer. In Proc. 2
. International Symposium on
Object-Oriented Real Time Distributed Computing,
IEEE Computer Society, Saint Malo, France.
Pankratius, V., & Stucky, W., 2005. A Formal Foundation
for Workflow Composition, Workflow View
Definition, and Workflow Normalization based on
Petri Nets. The 2nd. Asia-Pacific Conference on
Conceptual Modeling (APCCM 2005).
Schiefer, J., List, B., & Bruckner, R.M., 2003. Process
Data Store: A Real-Time Data Store for Monitoring
Business Processes, Lecture Notes in Computer
Science, Database and Expert Systems Applications
Sheth, A.P., van der Aalst, W.M.P., & Arpinar, I.B., 1999.
Processes Driving the Networked Economy: Process
Portals, Process Vortex, and Dynamically Trading
Processes. IEEE Concurrency, 7(3): 18-31.
Tjoa, A.M., List, B., Schiefer, J., & Quirchmayr, G., 2003.
The Process Warehouse – A Data Warehouse
Approach for Business Process Management.
Intelligent Management in the Knowledge Economy,
Edward Elgar Publishing, p. 112-126.
Totok, A., & Jaworski, R., 1998. Modelling of
Multidimensional Data Structures with ADAPT,
Technical Report, Technical University of
Braunschweig, ISBN 3-930166-92-5, 1998 (in
White, C., 2003. Building the Real-Time Enterprise,
TDWI Report Series, A101 communications
Publication. The Data Warehousing Institute.