Over the last few years, BI has proved to be a competitive differentiator for
enterprises that relied on BI tools for decision making. BI helps an enterprise
to improve its business performance by providing it the ability to achieve their
mission objectives by making smarter decisions at every functional level of the
business. Though the landscape of BI matured over years, it has come of age from
simple reporting and dash-boarding capabilities to incorporate analytics tools
as well. Analytics tools make a difference for organizations, by giving them
insights into the facts presented by reporting tools and then supporting them in
taking actions based on those facts. With the meteoric increase in data in
organizations, enterprises can now make those data accessible and useful by
tapping into the power of BI technologies. Today, BI systems are emerging in
every aspect of enterprise, ranging from large scale enterprise BI environments
that are deployed to support thousands of people, to smaller departmental BI
systems that cater to the needs specific to that department. The maturity of BI
adoption can be seen in new industry verticals like telecoms, retail,
health-care and BPO sectors.
For an enterprise to successfully implement BI, it becomes imperative to have
proper information management architecture in place. BI tools have to fetch for
information coming from various data stores across departments in an enterprise.
A BI tool sits atop all the data stores as a common layer. However, there always
remains a gap, because BI solution and the information management system, which
incorporates data-warehousing come from different vendors. This gap could be in
terms of compatibility of a BI solution with certain type of underlying
data-store or because of the fact that organizations avoid handling different
vendors and prefer one who can provide a single solution stack. This has
resulted in many joint ventures between BI solution providers and enterprise
solution providers, like MicroStrategy with Sybase, or acquisitions of BI
providers by enterprise solution providers, like BusinessObjects by SAP to
mention a few. This way, an enterprise can have a single solution that would
manage its information and also provide business intelligence solution. Apart
from having a singular stack solution for their BI needs, enterprises also want
BI solution to cover their whole enterprise and provide them with timely
information in an easy-to-interpret way with functional analytics that can help
enterprise make better business decisions. Systems that can provide such
information from cross departments and in faster time limits, are the part of
next gen BI tools. Here are few of the trends that are taking BI to a newer
landscape.
|
BI on Cloud
Cloud computing is one of the hottest trends in the industry today. Many
solution providers who used to serve on Software-as-a-Service (SaaS) platform
are now exploring Cloud as a platform to offer their services. In
business-economic terms, SaaS is predominantly driven by the wish to deploy IT
resources quickly and flexibly, at a time when organizational structures and
product portfolios are increasingly modified to comply with changing market
conditions. Cloud due to its benefits of auto-scalability and elasticity becomes
the next logical step for many to move into.
|
Though companies like SAP offer BI solutions (BusinessObject) on SaaS
platform as On-Demand offering, the likes of Google and Amazon have cloud
computing hardware infrastructure already. It's interesting to see how BI
companies would foray into this new domain and use such Cloud platform to
deliver their products. Already there were announcements from companies like
JasperSoft and Talend that they would expand their services on Cloud
infrastructure. Pehtaho has recently released its Cloud based BI product based
on Amazon EC2 platform. However, how enterprises adopt this remains to be seen.
Creating reports was considered to be a complex IT job. We want the users |
One drawback in a Cloud is that you lose control over the data which is
extremely sensitive and crucial to the decision making for an organization. You
can't risk landing your confidential data in the Cloud. What if the data reaches
your competitor's hands? That depends on how the Cloud environment matures up as
a enabler technology platform for providing On-Demand services.
Open Source BI solutions are considered more affordable than proprietary BI solutions and its adoption may pick up with subscription-based model that avoids an initial large payment for the software license. Open source BI platforms are getting designed to appeal to developers, who can take the software and embed data analysis into applications they develop. But enterprises have the typical fear, uncertainty, and doubt (FUD) for open source solutions over intellectual property liability, quality, security, scalability and support. Still organizations where there is a specific need and a defined budget are going ahead and exploring open source options. Sudhanshu Jain, AVP & Chief |
BI coming out of Silos
For an enterprise, analysis and reporting of data have become an integral part
of its decision making processes and actions. However, each organization has a
number of departments and each department has a BI solution which is specific to
its requirements. For instance, finance department might have its own BI system
residing on the financial information data-store. This makes viable sense for
having department wise BI, which in essence are the packaged BI tools that comes
along with the application. But when it comes to a BI solution that pans through
each of the processes of an enterprise, then it becomes necessary for the BI
solution to access various data-stores of different enterprise applications like
ERP, or SCM and provide cross functional reporting and analysis ability. This
cross departmental reporting and information analysis capability can be achieved
through new tech like multi-source.
Operational BI
BI has always been viewed as a tool to be used by decision makers and
business managers. Given the complexities involved, IT professionals seem to be
the prime users of BI tools. However, this is changing now. BI users now cover a
broad gamut from C-level executives to IT professionals to front-line business
users within an enterprise. Now everyone is allowed to access organizational
information that can help him in his work and generate his own reports. BI when
integrated with operational processes can allow organizations to react to
changing business conditions. For example, while a sales executive makes a sales
pitch to his client, the integrated BI capability can alert him of past deals
with the client and his buying pattern.
The information is delivered in a dashboard like interface which is easier to
interpret. Apart from the pre-generated reports and queries, the user has the
capability to fire ad-hoc queries, and with analytics tools he can do analysis
of what-if scenarios and predictions. The BI user interfaces now provide not
just basic reports, but advanced functionalities of various types also. BI
dashboards and portals are becoming richer and flexible due to implementation of
Rich Internet Application (RIA) tools that incorporate technologies like Flash
and Ajax. This helps deliver performance improvements for an employee. And, due
to dynamic dashboard interfaces, BI usage has now become available to almost
every user in the enterprise, thus making BI more operational in use.
BI Tools (open source) provided with the DVD
|
BI breaks free of data warehouse
BI tools fetch information from transactional sources like data stores, data
warehouses or data marts i.e. structured information sources. So for the BI
architecture, data warehouse becomes imperative. This is where from precise
reports can be generated through SQL queries. And in case of unstructured
information, such as, documents, emails, memos, etc, the SQL queries cannot
work, as this is a different architecture.
The new generation of BI solutions combine functionality of both structured
as well as unstructured content to provide information access over wider content
types. This is made possible through data integration layer, where the
unstructured content is transformed and loaded into a format that fits better
into a traditional analytics environment. This is how access to information,
through a BI tool, can be extended to linked information residing in documents,
emails, memos, spreadsheets etc. Typical scenarios where BI tools can access
structured as well as structured contents could be a BPO company that comprises
several applications. The BI system could extract information from various
applications and even from within emails.
New tech in BI
BI tools are becoming more faster in processing complex queries and going
beyond a functional process to access information from across departments.
In-Memory and Multi Source are the two technologies that are behind this
capability.
Multi-Source: A BI system should contain integrated functionalities that
include a database, an ETL technology (extract, transform and load) to move data
into the database, and a BI platform to create reports. A BI solution for a
particular department should allow the user to collect and store data from Excel
or databases and do report designing and conduct analyses against the data. But
for a cross departmental BI to work, the requirement would be accessing many
databases at once. It will require technical resource to aggregate all the data
of the enterprise, BI would need from a myriad of databases into the BI common
data-store, which is the metadata layer. Metadata is the layer where the BI
tools interact with the underlying data layers, i.e. the information from
various data stores is translated into business terms by BI tool that forms the
metadata. It will be a tedious task to move data from operational systems of
various departments into enterprise's data mart. Multi-source rescues you from
the task of migrating various databases. Multi-source allows you to quickly
incorporate various operational databases into the enterprise BI model without
going through the effort of moving the data into the data mart or data
warehouse. Once the new databases are identified in the BI metadata model, their
physical storage location becomes completely transparent to both BI users and
report developers.
SpagoBI incorporates various analytics engines that can be managed by the administrator by selecting Resources > Engine Management from the menu bar. The properties of the engines can be configured by clicking on Details icon. |
GeoEngine analytics allow location or map based analytics. Here, by clicking on each state location in the map we can have the corresponding sales information breakup for that state in the other panes. |
In Memory: The query performance of a BI system is fundamentally governed and
limited by the query performance of the underlying databases. The BI solution
providers have been developing techniques for SQL optimizations to improve query
performance for all major database technologies. One of these is In-memory query
caching. In-memory offers a new architectural approach for increasing query
performance on a very broad basis. In-memory BI operates like a database, but
has the performance characteristics of a memory-based caching system as well.
For instance, MicroStrategy's In-memory technique takes advantage of the huge
addressable memory space and is now available on 64-bit computers to provide
high performance middle-tier databases that can respond directly to data
requests from reports, dashboards, and OLAP analyses. Since the new middle-tier
databases are stored in computer memory, they avoid disk access delays of
traditional databases. A typical Online Analytical Processing (OLAP) cube
exhibits a performance profile in which execution of query time ranges from
minutes to hours depending on the volume of data and the complexity of the
query. With in-memory caching, the OLAP is supplemented with dynamic caching
technology that saves report instances in memory so that the next user of the
same report can be served the result directly from memory, avoiding database
delays. This creates a performance profile with two response-time distributions
— one for the database queries and one for the cache-based responses and hence
speeding up the BI performance time.
Analytics in BI Tools
Analytics have become an integral part of Business Intelligence suites.
There are many open source BI solution providers, but not all of them have
analytics capabilities. Pehtaho, JasperSoft, SpagoBI are some of the few open
source BI solutions that have analytics as part of the package. SpagoBI is an
open-source BI tool that offers wide variety of analytical tools and in an
intuitive user interface. Let's take a look at SpagoBI's analytics tools.
OLAP engines create complex cubes that gives complete data veiwpoints. In this OLAP inventory report, we can define additional parameters and by drill down feature you can access the breakup of the valus pertaining to a parameter. |
KPI monitoring is a key feature of any BI tool. SpagoBI provides the same wherein you can monitor performance of various indicators and can also see the value trend for any particular KPI by clicking on the TimeLine icon. |
About SpagoBI
The open source BI solution not only offers just simple reporting
functionality but also complex analytics capability through interactive
dashboards. It satisfies the whole range of typical BI requirements for an
organization in terms of data management and analysis along with administration
and security features. SpagoBI covers the whole range of analytical needs of an
organization with solutions for OLAP analysis, geo-referenced analysis,
data-mining techniques, free inquiry through Query by Example (QbE) module. It
also has intuitive dashboards to monitor the business performance indicators (KPI).
It also supports publication of data, which can be used as a collaborative
process between involved peoples for that report.
SpagoBi is an integration platform of various sets of tools and thus offers
many engines for the same analytical area. This gives the user freedom to choose
any analytical engines for his analytics requirements. All the integrated
engines share the same model that regulate the presentation of the analytical
document, and hence the interface remains same for all engines. It also has
integrated support for important processes of extraction, transformation and
loading (ETL) of data into the repository.
Getting started
You can find SpagoBI along with few other open source BI tools (check table)
in the DVD . SpagoBI comes as a ready to install zipped package along with
pre-configured Apache Tomcat 6.0 application server. All that you require is
that the machine should have a JDK installed and the JAVA_HOME environment
variable is defined. You just need to extract the contents of the zipped file.
From the bin folder of the base installation, you can start the Tomcat server
through startup.bat. Before starting off the server, you also have to define the
roles for users and the admin of the SpagoBI by defining them in tomcat-users.xml
file. After starting the Tomcat server, navigate on your browser to http://localho
st:8080/Spa goBI. Login as administrator and configure the analytic engines that
you want to use for generating reports and also define the databases that will
be used as data source by the analytics engines. While as a user, you can access
the KPI models and reports that are configured by the administrator. Based on
those reports and models, the user can modify, drill down and even do data
mining for accessing further information from them.
SpagoBI comes with a demo database, and we used the same for some report
generation and analytics as follows: