Moving Forward with the On Demand Real-time Enterprise

74 Moving Forward with the On Demand Real-time Enterprise

3.1 Introduction

Implementing a real-time enterprise environment is no simple task. It will take

time, expense, and energy. But, having such an environment will pay huge

dividends in terms of business viability and competitive business advantage. And

one of the first things you need as you start your journey, is a good roadmap. In

terms of this initiative, that roadmap is actually an architecture.

The objective of this chapter is not to present you with such an architecture, but

rather discuss the drivers, requirements, and a framework for one. You will then

need develop your architecture based on your specific requirements.

From our perspective, there are three major drivers to be considered when

developing a real-time enterprise architecture. They are:

1. Data silos

2. Isolated decision making

3. Latency in the decision making process

The following is a brief discussion of those drivers.

Data silos

Much of the data available in an enterprise today is housed in what we call data

silos

. These are sources of data that have been collected and managed by

specific business functional areas. They are typically not integrated with other

data in the enterprise, and, in fact, may not even be shared with them. And the

data in these various silos is typically at differing levels of currency, and may well

even have differing data definitions (metadata) for the same business entities.

Having these data silos at differing levels of currency, and un-integrated, are

major reasons why there is inconsistency in reporting and data analysis in

businesses today.

Data from these diverse sources typically takes a long time to be consolidated

into a consistent, unified global view that allows the business analyst to discover

potential problems, take corrective action, identify new business opportunities, or

detect anomalies. In many cases, the massive amounts of data from

brick-and-mortar stores, Web sites, call centers, inventory, and CRM systems

are often analyzed independently from a business unit perspective, rather than

integrated to realize an enterprise-wide view.

This is typically because the systems have been developed independently. As

such, they likely do not communicate with each other. And the data from these

systems may be represented in different formats, stored in different databases

and captured in diverse applications.