We take the case of a multibillion-dollar automotive company that has business units spread across geographies. The automotive company is structured across geographies as strategic business units (SBUs). SBUs have a mix of both legacy as well as modern systems (e.g., ERP, SCM, and CRM). The systems of the automotive company across SBUs are as follows:

1.  North America

• Legacy systems (Mainframes)

2.  Europe

• Legacy + SCM

3.  Asia

• ERP + SCM + CRM

4.  South America

• Java based + Oracle RDBMS

The automotive company wants to offer its CEO a top-level view of the performance of its SBUs. The KPIs required are:

1.  Revenues by SBU

• Drill down (break up by product categories e.g., trucks, cars)
  - Drill down (break up by brands)

2.  Revenues - target vs. actual, by SBU

• Drill down (break up by product categories)
  - Drill down (break up by cars)

3.  Revenue forecast by SBU

• Drill down (break-up by product categories)
  - Drill down (break-up by cars)

4.  Sales volumes by SBU

• Drill down (break up by product categories)
  - Drill down (break up by cars)

5.  Production volumes by SBU

• Drill down (break up by manufacturing plants)
  - Drill down (break up by cars)

6.  Top dealers by SBU

• Drill down (break up by product categories)
  - Drill down (break up by cars)

7.  Profit margins by SBU

• Drill down (break up by product categories)

Current Enterprise Architecture
The current high-level architecture of the systems is illustrated in Figure 1. This diagram highlights the scatter of critical data across geographies. Given the nature of the problem requirements and the diversity of the data sources involved, it seems that the management of the data sources and the centralization of common data across lines of business and geographies would prove to be very important in fulfilling the business requirements. Some of the specific issues and challenges in this regard are listed below.

• Heterogeneous applications: Core data is required by the dashboard as well as by other enterprise applications. In order to be consistent, the data must be integrated and disseminated using open standards. All consumers of data must obtain data from the same source. Apart from this, the dashboard would require leveraging logic from several applications that are written using different technologies. These disparate applications must be integrated to present consistent information to the dashboard.
• Incomplete view of data: The scatter of critical data, e.g., customer data across multiple databases in various businesses and geographies creates a serious issue with respect to the identification of key customer attributes.
• Complex systems logic: The spread of multiple systems and databases will result in business applications having to embed the logic associated with the retrieval/update of multiple data elements residing in different databases.
• Financial impact: Depending on the manner in which specific technologies have evolved in a given LOB (Line of Business), the data access techniques may vary between applications. This can lead to higher costs incurred for integration across lines of business due to the difference in technical abilities between lines of business.
• Problems due to multiple channels of update: As discussed above, the core business data in a specific database may be updated by a number of mechanisms such as direct update through the business applications dedicated for that line of business, indirect update through synchronization, and update through other channels.
• User experience issues: Such a situation may arise when a user requests a change in personal information over the phone or IVR and attempts to look up the same data online. Latencies in the update of the data that has been sent over the phone may cause incorrect data to be displayed.
• Resource contention: This issue can arise due to heavy loads on any of the databases arising during a bulk update situation.
• Latencies in data retrieval: The spread of critical data across multiple databases and the associated redundancy may cause additional latencies for certain applications. This can happen due to the bulk update scenario illustrated above.

The proposed architecture for the dashboard consists of two main tiers: the data access tier and the enterprise dashboard tier. The enterprise dashboard provides an aggregated view of the enterprise information collected from varied data sources that are geographically dispersed.