Moving Forward with the On Demand Real-time Enterprise

Chapter 1. Introduction 23

You should also look at the information accelerator layer, that includes such

added value as data models, profile models, schemas and policies for specific

industries. There are over 400 of these that are available, not just to IBM clients,

but also to IBM Business Partners.

Information on demand is really all about overcoming information complexity,

and about information enabling your business applications and business

processes. It enables you to optimize your information environment through IT

services management, and along with that, reduces the hardware cost through

autonomic and grid computing capabilities. The net result is that it builds added

value for you.

1.2.2 Moving to real-time

So what are some of the changes that moving to a real-time enterprise brings

with it. We have summarized some of them in the following list.

򐂰 Event driven: Rather than relying only on analysts to analyze you data and

take an action, processes need to be event driven. The events trigger action,

such as the development of real-time business intelligence through the use of

real-time analytics. Embedded, or in-line, analytics interact directly with the

business processes, triggered by alerts and based, for example, on key

performance indicators and business rules. There are also embedded data

mining techniques that can be directly invoked, as well as the direct

enablement of such capabilities as OLAP and the creation of components

such as data cubes for analysis.

򐂰 Rules-based: To enable event processing, and particularly an automated

response to an event action, there is a need for a rules-based environment.

These business rules can be accessed by analytic applications, for example,

for automated decision-making action. The rules base must be tightly

integrated with the business processes, and kept up to date. So, the rules

should be flexibly implemented so they can be easily changed or updated as

the business requirements change.

򐂰 Updating the data warehouse: Updates to the data warehouse have

historically been performed primarily in a batch mode. That is, the updates

are typically performed on a scheduled basis, such as nightly or weekly, with

a batch of transactions that have been collected over some period of time.

The requirement now is to perform those data warehouse updates in more of

a continuous streaming mode. That is, updates may need to flow to the data

warehouse in more of a continuous manner, soon after the transactions

occur. Of course, this will require changes to the business processes and the

IT support processes. It will also require changes to the ETL (extract,

transform, and load) environment.