Moving Forward with the On Demand Real-time Enterprise

Chapter 3. Architectural considerations 79

The IT organization

IT organizations today are aware of the value propositions provided by

information management technologies. The focus of most discussions has

shifted beyond information management itself to the larger issues surrounding

the development of an overall architecture for implementing a real-time

enterprise.

Data warehousing technologies remain a vital foundation for any strategic

approach to leveraging all of an organizations information assets. It establishes

consistent, predictable levels of data quality, consistency, and stability.

However, the scope of business intelligence architectures has expanded beyond

traditional data warehousing disciplines such as data extract, transform and load

(ETL) and reporting. Today, business intelligence requires information

on

demand

. This may include actionable business insights from combining functions

such as data mining and multidimensional data analysis, with advanced

statistical and analytical functions in a low latency real-time, integrated

environment. OLAP and data mining, in particular, have brought with them

specialized tools, APIs, and data structures. And also purpose-built engines,

spanning all levels of the system infrastructure, from front-end client tools, to

middle-tier server cache, to back-end data warehouse.

The multi-tier, enterprise-wide nature of BI tools and infrastructure demands a

high-level architectural model, a framework for BI. A BI framework provides an

overall conceptual model for understanding, planning and managing this

complex topology. The BI framework should express the strategy, ideally in a

way that differentiates the particular approach from others. It should also be

capable of enlightening customers regarding what to expect in terms of future

direction, technology and interfaces. And, it should guide system architects and

integrators in selecting, designing and deploying BI tools and applications.

Data governance and documentation

Data governance is required in any organization to ensure consistent data quality

and meaning. Having good standards for formatting, key matching, and

distributed maintenance is still a big issue in many organizations.

As systems evolve and data changes, perhaps simply moving to a new operating

environment will result in changes to the meaning of the data. Therefore, good

documentation should follow and be kept updated.

3.1.3 Creating an information infrastructure

In the following sections we illustrate solutions that will provide resolutions to the

information management challenges. Here we look at how to enable a business