Business Intelligence Architecture

Rome wasn't built in a day, so does a business intelligence system of any organization. It requires lot of strategies and efforts to build such a solution.

To start with the effort of building business intelligence system, one needs a framework comprising of best practices, policies, and standards. Business intelligence architecture, by providing this framework, ensures that the development efforts of multiple projects fit neatly together as a cohesive whole to achieve desired BI system.

Let us briefly explore the architecture of business intelligence system.

Why we need business intelligence architecture?

Much before an organization starts adopting a business intelligence architecture, there are series of indicators which accelerate the case for building a BI system. There are many key factors, but important ones include:

  • Backlog of business requests: IT department will be in lot of pressure to fulfill the report requests from various business users.
  • Need for self-service BI: Business users are struck as they need to depend on IT for every small piece of information. This hinders their decision making process and comes as a bottleneck for smooth operation.
  • Messed up IT system: Silos of data, different data formats, disparate data and applications - these will form a complex IT system and will build a strong case for a stronger BI infrastructure.
  • Cost: Cost of maintaining information silos and feeding to huge number of IT resources for even small set of data is not a good thing for organization. They become huge cost centers

So, the organizations end up in building a business intelligence architecture that will seek to help organizations and businesses make better decisions. A solid architecture will help to structure the process of improving your business intelligence and helps you to implement your Business Intelligence strategy in a very cost effective way.

BI architecture, among other elements, often includes both structured and unstructured data. This data comes from both internal and external sources and are transformed from raw transaction data into logical information. Let us explore the major components of BI architecture in the next discussion.

Components of business intelligence archicture

One mistake that top leaders of many organization make is think of their BI system as equivalant to front-end BI tools being used. Then there is another set of technical geeks who make lot of dicussion about a business intelligence architecture around some fancy jargons without actually giving due importance to what exactly constructs BI architecture.

Below list will covers major part of the business intelligence architecture - well, there are few more but as said, these are important ones.

  • Source systems
  • ETL process
  • Eata modelling
  • Data warehouse
  • Enterprise information management (EIM)
  • Appliance systems
  • Tools and technologies

Source systems - transaction processing systems

This is the starting point for any BI initiative. Organization data is first created in these databases. Important point is -
If you do not capture the data in the operational system, you can’t analyze it.

Operational systems(OLTP) forms bulk of the data needed for data warehouse. In addition to that, some time source systems will include data from secondary sources such market data, benchmarking data etc. Business intelligence architecture should address all these various data sources which are of different format, and standards.

ETL process

A details ETL process in a data warehouse considers many things. In summary, in an ETL process data will be extracted from the operational systems and loaded into a data warehous. ETL, which stands for Extract Transform Load, is usually done using custom solutions available in the market. IBM Websphere Data Stage, Oracle Data Integrator, Ab Initio, and Microsoft Integration Services are examples of such tools.

Data modeling

Data modeling will help to address what exactly is needed from data sources and format of the data and how it will be related to other data elements. It is not feasible to extract everything from a source system as that will cost a lot. Data modeling will help to organize the data and therefore will minimize cost of storage replication, and effort needed to build a data warehouse.

Data warehouse

Warehouse will have data extracted from various operational systems, transformed to make the data consistent, and loaded for analysis. Some people argue that a separate data warehouse is not a necessary - we will not debate it here. But let is assume that it is a necessity for all organization who want to implement a BI solution.

A data warehouse will help to achieve cross-functional analysis, summarized data, and maintaining one version of truth across the enterprise.

Enterprise information management (EIM)

EMI is another BI jargon which may stump some beginners. The term usually refers to ETL tools, data modeling tools, data quality, data profiling, metadata management, and master data management. Without going deep into each of these, let us understand EIM as a way to achieve optimal use of any set of information within the organization.

BI hardware

It is important to make decision on what hardware is required to maintain a high performance and scalable BI system. Apart from server configuratios, we have data warehouse appliances combine the server, the database, and the data storage into one system. Netezza and DATAllegro are some well known appliances in the market.

Tools and technologies

Another important component of business intelligence architecture is what tools and technologies to implement. It is not just the front-end UI tools, but the tools used for EIM as well. There are cloud solutions, SaaS model, many full fledged BI solutions(such as MSBI, Oracle BI suites, Microstrategy and more) to choose from. BI framework should guidelines to make decision on what is required for the need of the organization.