Thursday 16 May 2013

Data Integration Applications And Techniques

By Peggie K. Lambert


The basic concept behind data integration is that it combines data in several sources and puts it forward to an application or user as coming from a single source. This can be achieved using different methods. The range of applications, including marketing, corporate mergers and research, is just as broad.

The main reason to integrate sources has to be related to business process enhancement, as opposed to IT management. This means the integration should bring new capabilities or solve some problem, instead of simply being a cheaper or better way for the company to store its information. Many such business-enhancing integrated systems have been created and are commonly used by companies in many different industries.

For example, a company that has separate databases for their marketing/sales and service departments may want to integrate all of it into a centralized repository. Otherwise known as a CRM system, this allows the marketing and sales people to target existing customers based on information collected from the other departments. Two companies entering into a merger or acquisition will likewise need to integrate their respective systems.

It is possible to accomplish this process on both the application and middleware layers. It may also be done by physical warehousing in a completely new system, or through virtual integration where no new repository is actually created. Let's consider these choices one by one, so that it becomes clear how each one works and where it can be used.

Many applications are capable of integrating information from multiple sources. These applications will therefore not require to be plugged in with a separately designed integration tool. Some systems have this integrating capability built into middleware layers, which can access all the sources and retrieve the required information for applications upon demand.

Virtual integration doesn't actually involve creating a separate tool or software solution. It merely defines a set of views that can help users pull together information from disparate sources. For instance, if an employee wants to view one customer's profile, the pre-defined query simply calls in all the records for this unique customer ID from every department and presents it together as a single unified view.

Physical warehousing involves creating a separate storage system which collects information from all sources. This is an enterprise-level process, where information flows in from all of a company's locations and departments into a single data center. This is a centralized system which facilitates enterprise-wide reporting and analysis. It also helps integrate applications by providing all of them access to the pooled data.

The exact data integration techniques to be applied and the project's scope need to be decided very carefully. The business benefits and the type and quantity of sources are all relevant factors. The implementation cost has to be considered, along with backup systems and other security issues. Other aspects that impact the project or may be impacted themselves include synchronization, migration and master data management or MDM.




About the Author:



No comments:

Post a Comment