Moving Forward with the On Demand Real-time Enterprise

6 Moving Forward with the On Demand Real-time Enterprise

1.2 The on demand real-time enterprise - an overview

An on demand business needs to extend its information management strategy

by transforming its business processes and changing how the company uses

information to drive insight. For example, the integration of heterogeneous data

sources is fast becoming a basic requirement to satisfy the needs of the

business. This will enable you to adapt to the changing environment, and

leverage your data for a business advantage. It will also provide employees and

applications with easy access to an integrated source of enterprise information.

But do we really need a real-time enterprise? Do we need access to all the data

in the enterprise instantaneously, or at least very quickly?

Unfortunately the answers are either yes, no, maybe, and sometimes. And then

the ever popular follow-up of

it all depends. And one of the things upon which it

depends is your definition of a real-time enterprise.

Instantaneous access or current data?

But, hold on just a minute. We have already arrived at another point of

contention. Real-time does not simply mean having access to the data

instantaneously after some event. Real-time data is current data relative to a

particular business event. So real-time is relative to the completion of the event,

rather than the speed of access to it.

For example, say an event only happens once a day. The data about that event

can still be considered real-time anytime within a 24-hour period, because it is

still the most current data. However, we agree that the ability to have that data as

quickly as possible after the event will make it even more valuable. So, we can

agree that this is about semantics. The message is, do not be overly concerned

about the semantics. Be more concerned about the functionality, practicality, and

cost justification of the environment.

More than data access

You should understand that to become a real-time enterprise requires more than

simply having instant access to current data. You must implement support to

make sure current data is really current!

Basically, a real-time enterprise is similar to what we at IBM have referred to as

an on demand enterprise. Here is another point of reference.

Attention: An on demand enterprise can be thought of as one whose

business processes—integrated end-to-end across the enterprise and with

key partners, suppliers and customers—can respond with the appropriate

speed to any customer demand, market opportunity, or external threat.

Chapter 1. Introduction 7

However, that description implies integration and action. Integration is all about

enabling people and applications to interact with other people and applications.

That lets you leverage information, based on end-to-end business processes

within and beyond the enterprise. It requires business flexibility, agility and IT

simplification. Therefore, using the terms in combination is very appropriate. That

is, we now use the term

on demand real-time enterprise. A high-level view of

such an on demand real-time enterprise is depicted in Figure 1-1.

Figure 1-1 On Demand Real-time Enterprise - an example

The enterprise data warehouse (EDW) can be continuously updated by

processes such as extract, transform, and load (ETL), extract, load, and

transform (ELT), messages that are passed via queues, replicated data from

tables or queues, and Web services. The data can be both structured data and

un-structured content. However, this activity does not impact the continuous

access to the data warehouse by such processes as concurrent queries, analytic

applications, data mining, and corporate dashboards.

When we talk about the EDW, we are typically talking about an enterprise data

warehousing

environment. That is because the environment consists of the

EDW, operational data store (ODS), as well as physical and virtual data marts.

Including the ODS in the data warehousing environment enables access to more

current data more quickly, particularly if it happens that the data warehouse is

updated by one or more batch processes rather than continuously.

The following list contains a few more things you will need to consider as you

move towards an on demand real-time enterprise:

Parallel ETL

Engines

Message

Delivery by

Queues

Replication

Web

Services

Personalization,

data mining,

business rules,

campaigns

Alerts, triggers,

KPIs, analytics

Corporate

Dashboards

Concurrent

Queries

Consumers

Continuous

Data

Loading

Enterprise

Data

Warehouse

Content