Description
Business Intellligence should be described as a way of managing our company more than a set of functionalities in a computer software.
Abstract— Business Intellligence should be described as a
way of managing our company more than a set of functionalities
in a computer software. Acquiring a real profit requires
enterprise management to understand the value of the data and
the way data describe business processes. Being aware of the
business and measuring its performance we are able to improve
the processes and make whole the business more effective. To
achieve business improvement we require efficient Business
Intelligence system as a combination of a software, hardware,
the communication infrastructure and services regarding data
preparation, integration and delivery to the system.
In this paper author considers if the service oriented
approach and cloud computing can make BI implementation
more efficient.
I. INTRODUCTION
USINESS INTELLIGENCE for sure is the way for the
enterprise to make the business more efficient. BI
systems are fed by a big volumes of data coming from the
transactional systems working inside company. That is why
the BI systems are most often on-premise software which
mean they are hosted on the servers inside the building the
company is located.
Another way of using software is called software as a service
or cloud computing which can bring the significant benefits
for a company. Are these benefits strong enough to make BI
implementation in cloud environment more efficient than on-
premise?
II. BI IMPLEMENTATION
Different definitions show that the components of business
intelligence software are inter ailia[1]:
? Data warehousing
? Multidimensional analysis, for example OLAP
? Data mining
? Business analysis
? Visualization
? Querying, reporting and charting (including just-in-
time and agent-based alerts)
? Geospatial analysis
The software is supplied by the following categories of
data:
? Operational data (regarding financial, logistics, sales,
orders, personnel, billing)
? Private data (mainly spreadsheets prepared by
business analysts, knowledge workers, statisticians
and managers regarding analysis and reports)
? External data (purchased from vendors specializing
in collecting industry-specific information such as:
Health care statistics, customer profile information,
customer credit reports, trends, currency
fluctuations, stock prices demographic data and
many more)
Operational data in most cases come from the
transactional systems hosted inside the company. They are
the biggest data volumes from all mentioned categories.
Operational data is probably the most important data
category because the data describe directly our business.
In a current situation most of BI systems are implemented
basing on the following assumptions:
? BI system is located in the same LAN as the main
transactional systems, because it collects the huge
amount of data from transactional systems,
? Data loading is performed once a day at the time
when users do not access the systems,
? The internal structure of the data warehouse is unique
for every implementation (because every enterprise
has unique combination of transactional systems
that they use).
? The user interfaces are usually typical for business
analytics, reporting, visualizations, charting and
manager dashboards
? BI system needs a lot of resources, mainly: disk
storage because of the volumes of data, network
capacity during data loading, processor and
memory during data processing.
We have to admit, that BI is evolving. The new
possibilities of use appear what implies new challenges
for Business Intelligence. We can find among them
[3],[8]:
? Big Data analytics
B
Business Intelligence as a service in a cloud environment
Maciej Pondel
Wroclaw University of Economics
ul. Komandorska 118/120,
53-345 Wroc?aw, Poland
Email: [email protected]
Proceedings of the 2013 Federated Conference on
Computer Science and Information Systems pp. 1269–1271
978-1-4673-4471-5/$25.00 c 2013, IEEE 1269
? Mobile BI
? In memory BI
? Self service BI
? Consumerization of Enterprise Software
III. SERVICE ORIENTED ARCHITECTURE
Service oriented architecture of software is not a new
technology [4]. It is rather the idea of creating software as a
set of modules called services that collectively provide the
complete functionality of a large software application. The
services should cooperate by exchanging data and
information with other services without any human
interaction. The services should be treated as black boxes
with precisely defined input parameters and output results.
For the architect of enterprise application consisted of
services the algorithms implemented inside the service are
not so important. The services can have various types. We
can have:
? business service – simple IT component
responsible for a part of business,
? Web Service – as a method of communication
between 2 programs based on World Wide Web
protocol
? IT Service – as a business process of supplying
benefits to the recipient by a supplier
In order to build efficient SOA solutions the services must
meet following requirements:
? Interoperability between different programming
languages, systems that allow integration of
services
? Federation of resources that allows transparently
mapping multiple autonomous resources to be
treated by users as one federated resource
Cloud is becoming more and more popular environment for
hosting business applications. All greatest IT vendors in the
world provide their software as a service available in the
cloud. It means the systems are hosted in the vendor’s
environment on vendor’s severs and the systems are
available by Internet connection. To the most popular IT
systems provided in cloud belong:
? Enterprise email systems together with tools
regarding workgroup like shared calendars,
resources reservation and applications improving
employees’ productivity,
? Document management systems,
? Content management systems,
? Databases (like Google Cloud SQL or Windows
Azure SQL),
? Business applications like CRM, ERP,
? Application hosting services with programmers API
allowing customer to build his own software and
host it in the cloud.
The main assumption of cloud services according to
Forrester’s definition is supplying standardized IT capacity
over Internet in a pay-per-use and self-service way [7]. We
can understand, that it is suitable way of using IT for all
systems that works in a standard way. It we have to use
dedicated software the cloud can be too limited for us.
According to the Internet technologies – cloud is supporting
the standards prevalent in mobile devices. Most of the
standardized cloud services can be run at any devices with
any operating systems.
What are the main benefits or using cloud services in
business[6]?
? It lowers the costs of entry for smaller companies
who are trying to use the same software and
technologies as the big corporations.
? If the IT solution needs a large amount of
computing power for relatively short time the
cloud can provide us the resources dynamically
what is financially available even for small
companies.
? It provides immediate access to hardware
resources without any upfront investments. It
shortens the time to market for many IT
solutions.
? Even for large enterprises cloud can allow to scale
their applications in a simpler way. The
company has an easy access to the new
computer resources whenever then are
necessary.
? Cloud computing requires payments only for the
resources the company really utilizes. In many
cases it is more effective financially than
investing the money on the start.
When we compare the characteristics of the cloud
environment and the benefits with the assumptions of BI
implementations and challenges of BI we can observe that
they match each other. Many of those benefits concerning
cloud are suitable for Business Intelligence solutions. The
features of BI Systems and cloud are summarized in table 1.
As we can observe there are some features that do not fit
each other. The basic problem concern the huge data transfer
from transactional systems hosted mainly on-premise to the
solution kept in cloud. The second issue is about dedicated
integration procedures and unique data warehouse structure
what do not match the standardization of the cloud solutions.
We have to admit that those issues will not occur when we
have transactional systems hosted also in the cloud – the
cloud BI system is a very natural solution because:
? data during the loading stay in the same environment
and do not overload the network connection
? the procedures of integration will be typical because
the cloud transactional systems must be
implemented in a common way
1270 PROCEEDINGS OF THE FEDCSIS. KRAK
´
OW, 2013
If we have most common architecture of transactional
systems and they are hosted on-premise we have to elaborate
the solution for the indicated issues.
One of the solutions may be an adoption of a hybrid BI
environment consisting of:
? the data integration layer which is stored on
premise
? the business logic layer stored in the cloud
IV. CONCLUSION
This paper presents the idea of cloud BI system. There are
some cases that may inhibit such model of BI software
functioning, but author presents how to manage them and
what advantages gives combining features of a cloud
environment with user expectations regarding BI systems.
REFERENCES
[1] Moss L., Atre S. Business Intelligence Roadmap, Pearson Education,
Boston 2003
[2] Wu L., Barash G., Bartolini C. „A Service-oriented Architecture for
Business Intelligence”, Service-Oriented Computing and
Applications, 2007. SOCA '07. IEEE Computer Society, Los
Alamitos 2007
[3]http://bi.pl/publications/art/99-trendy-w-business-intelligence-z-
czym-to-sie-je-jak-czytac-menu-i-co-zamowic#section_section-375
[4] ?agowski L, „SOA – Ideologia nie technologia”, XV Konferencja
PLOUG, Ko?cielisko, 2009
[5] Velte, Anthony T. (2010). Cloud Computing: A Practical Approach.
McGraw Hill. ISBN 978-0-07-162694-1.
[6] Marston S, Li Z, Bandyopadhyay S, Zhang J, Ghalsasi A. Cloud
computing: the business perspective. Decis Support Syst. 2011
[7] Lecznar M., Patig S. Cloud Computing providers: Characteristics and
recommendations in: Babin G, E-Technologies: Transformation in a
Connected World: 5th International Conference, MCETECH 2011,
Springer Berlin-Heidelberg 2011
[8] Howson C., 7 Top Business Intelligence Trends For 2013:http://www.informationweek.com/software/business-intelligence/7-
top-business-intelligence-trends-for-2/240146994
TABLE I.
BI SYSTEMS AND CLOUD CHARACTERISTICS
BI Characteristics and challenges\
cloud features
Standardized
solutions
Internet
technologies
Pay-for-use Dynamic
allocation of
resources
Hosting in
the vendor
environment
Huge data transfer from transactional
systems
-
Data loading once a day +
Unique data warehouse structure and
unique integration
-
Typical user interfaces +
Needs a lot of resources in various part of
time
+ +
Big data analysis
Mobile BI + +
In memory BI
Self service BI +
MACIEJ PONDEL: BUSINESS INTELLIGENCE AS A SERVICE IN A CLOUD ENVIRONMENT 1271
doc_328633757.pdf
Business Intellligence should be described as a way of managing our company more than a set of functionalities in a computer software.
Abstract— Business Intellligence should be described as a
way of managing our company more than a set of functionalities
in a computer software. Acquiring a real profit requires
enterprise management to understand the value of the data and
the way data describe business processes. Being aware of the
business and measuring its performance we are able to improve
the processes and make whole the business more effective. To
achieve business improvement we require efficient Business
Intelligence system as a combination of a software, hardware,
the communication infrastructure and services regarding data
preparation, integration and delivery to the system.
In this paper author considers if the service oriented
approach and cloud computing can make BI implementation
more efficient.
I. INTRODUCTION
USINESS INTELLIGENCE for sure is the way for the
enterprise to make the business more efficient. BI
systems are fed by a big volumes of data coming from the
transactional systems working inside company. That is why
the BI systems are most often on-premise software which
mean they are hosted on the servers inside the building the
company is located.
Another way of using software is called software as a service
or cloud computing which can bring the significant benefits
for a company. Are these benefits strong enough to make BI
implementation in cloud environment more efficient than on-
premise?
II. BI IMPLEMENTATION
Different definitions show that the components of business
intelligence software are inter ailia[1]:
? Data warehousing
? Multidimensional analysis, for example OLAP
? Data mining
? Business analysis
? Visualization
? Querying, reporting and charting (including just-in-
time and agent-based alerts)
? Geospatial analysis
The software is supplied by the following categories of
data:
? Operational data (regarding financial, logistics, sales,
orders, personnel, billing)
? Private data (mainly spreadsheets prepared by
business analysts, knowledge workers, statisticians
and managers regarding analysis and reports)
? External data (purchased from vendors specializing
in collecting industry-specific information such as:
Health care statistics, customer profile information,
customer credit reports, trends, currency
fluctuations, stock prices demographic data and
many more)
Operational data in most cases come from the
transactional systems hosted inside the company. They are
the biggest data volumes from all mentioned categories.
Operational data is probably the most important data
category because the data describe directly our business.
In a current situation most of BI systems are implemented
basing on the following assumptions:
? BI system is located in the same LAN as the main
transactional systems, because it collects the huge
amount of data from transactional systems,
? Data loading is performed once a day at the time
when users do not access the systems,
? The internal structure of the data warehouse is unique
for every implementation (because every enterprise
has unique combination of transactional systems
that they use).
? The user interfaces are usually typical for business
analytics, reporting, visualizations, charting and
manager dashboards
? BI system needs a lot of resources, mainly: disk
storage because of the volumes of data, network
capacity during data loading, processor and
memory during data processing.
We have to admit, that BI is evolving. The new
possibilities of use appear what implies new challenges
for Business Intelligence. We can find among them
[3],[8]:
? Big Data analytics
B
Business Intelligence as a service in a cloud environment
Maciej Pondel
Wroclaw University of Economics
ul. Komandorska 118/120,
53-345 Wroc?aw, Poland
Email: [email protected]
Proceedings of the 2013 Federated Conference on
Computer Science and Information Systems pp. 1269–1271
978-1-4673-4471-5/$25.00 c 2013, IEEE 1269
? Mobile BI
? In memory BI
? Self service BI
? Consumerization of Enterprise Software
III. SERVICE ORIENTED ARCHITECTURE
Service oriented architecture of software is not a new
technology [4]. It is rather the idea of creating software as a
set of modules called services that collectively provide the
complete functionality of a large software application. The
services should cooperate by exchanging data and
information with other services without any human
interaction. The services should be treated as black boxes
with precisely defined input parameters and output results.
For the architect of enterprise application consisted of
services the algorithms implemented inside the service are
not so important. The services can have various types. We
can have:
? business service – simple IT component
responsible for a part of business,
? Web Service – as a method of communication
between 2 programs based on World Wide Web
protocol
? IT Service – as a business process of supplying
benefits to the recipient by a supplier
In order to build efficient SOA solutions the services must
meet following requirements:
? Interoperability between different programming
languages, systems that allow integration of
services
? Federation of resources that allows transparently
mapping multiple autonomous resources to be
treated by users as one federated resource
Cloud is becoming more and more popular environment for
hosting business applications. All greatest IT vendors in the
world provide their software as a service available in the
cloud. It means the systems are hosted in the vendor’s
environment on vendor’s severs and the systems are
available by Internet connection. To the most popular IT
systems provided in cloud belong:
? Enterprise email systems together with tools
regarding workgroup like shared calendars,
resources reservation and applications improving
employees’ productivity,
? Document management systems,
? Content management systems,
? Databases (like Google Cloud SQL or Windows
Azure SQL),
? Business applications like CRM, ERP,
? Application hosting services with programmers API
allowing customer to build his own software and
host it in the cloud.
The main assumption of cloud services according to
Forrester’s definition is supplying standardized IT capacity
over Internet in a pay-per-use and self-service way [7]. We
can understand, that it is suitable way of using IT for all
systems that works in a standard way. It we have to use
dedicated software the cloud can be too limited for us.
According to the Internet technologies – cloud is supporting
the standards prevalent in mobile devices. Most of the
standardized cloud services can be run at any devices with
any operating systems.
What are the main benefits or using cloud services in
business[6]?
? It lowers the costs of entry for smaller companies
who are trying to use the same software and
technologies as the big corporations.
? If the IT solution needs a large amount of
computing power for relatively short time the
cloud can provide us the resources dynamically
what is financially available even for small
companies.
? It provides immediate access to hardware
resources without any upfront investments. It
shortens the time to market for many IT
solutions.
? Even for large enterprises cloud can allow to scale
their applications in a simpler way. The
company has an easy access to the new
computer resources whenever then are
necessary.
? Cloud computing requires payments only for the
resources the company really utilizes. In many
cases it is more effective financially than
investing the money on the start.
When we compare the characteristics of the cloud
environment and the benefits with the assumptions of BI
implementations and challenges of BI we can observe that
they match each other. Many of those benefits concerning
cloud are suitable for Business Intelligence solutions. The
features of BI Systems and cloud are summarized in table 1.
As we can observe there are some features that do not fit
each other. The basic problem concern the huge data transfer
from transactional systems hosted mainly on-premise to the
solution kept in cloud. The second issue is about dedicated
integration procedures and unique data warehouse structure
what do not match the standardization of the cloud solutions.
We have to admit that those issues will not occur when we
have transactional systems hosted also in the cloud – the
cloud BI system is a very natural solution because:
? data during the loading stay in the same environment
and do not overload the network connection
? the procedures of integration will be typical because
the cloud transactional systems must be
implemented in a common way
1270 PROCEEDINGS OF THE FEDCSIS. KRAK
´
OW, 2013
If we have most common architecture of transactional
systems and they are hosted on-premise we have to elaborate
the solution for the indicated issues.
One of the solutions may be an adoption of a hybrid BI
environment consisting of:
? the data integration layer which is stored on
premise
? the business logic layer stored in the cloud
IV. CONCLUSION
This paper presents the idea of cloud BI system. There are
some cases that may inhibit such model of BI software
functioning, but author presents how to manage them and
what advantages gives combining features of a cloud
environment with user expectations regarding BI systems.
REFERENCES
[1] Moss L., Atre S. Business Intelligence Roadmap, Pearson Education,
Boston 2003
[2] Wu L., Barash G., Bartolini C. „A Service-oriented Architecture for
Business Intelligence”, Service-Oriented Computing and
Applications, 2007. SOCA '07. IEEE Computer Society, Los
Alamitos 2007
[3]http://bi.pl/publications/art/99-trendy-w-business-intelligence-z-
czym-to-sie-je-jak-czytac-menu-i-co-zamowic#section_section-375
[4] ?agowski L, „SOA – Ideologia nie technologia”, XV Konferencja
PLOUG, Ko?cielisko, 2009
[5] Velte, Anthony T. (2010). Cloud Computing: A Practical Approach.
McGraw Hill. ISBN 978-0-07-162694-1.
[6] Marston S, Li Z, Bandyopadhyay S, Zhang J, Ghalsasi A. Cloud
computing: the business perspective. Decis Support Syst. 2011
[7] Lecznar M., Patig S. Cloud Computing providers: Characteristics and
recommendations in: Babin G, E-Technologies: Transformation in a
Connected World: 5th International Conference, MCETECH 2011,
Springer Berlin-Heidelberg 2011
[8] Howson C., 7 Top Business Intelligence Trends For 2013:http://www.informationweek.com/software/business-intelligence/7-
top-business-intelligence-trends-for-2/240146994
TABLE I.
BI SYSTEMS AND CLOUD CHARACTERISTICS
BI Characteristics and challenges\
cloud features
Standardized
solutions
Internet
technologies
Pay-for-use Dynamic
allocation of
resources
Hosting in
the vendor
environment
Huge data transfer from transactional
systems
-
Data loading once a day +
Unique data warehouse structure and
unique integration
-
Typical user interfaces +
Needs a lot of resources in various part of
time
+ +
Big data analysis
Mobile BI + +
In memory BI
Self service BI +
MACIEJ PONDEL: BUSINESS INTELLIGENCE AS A SERVICE IN A CLOUD ENVIRONMENT 1271
doc_328633757.pdf