Businesses demand immediate, ready access to real-time data integrated from disparate sources. Software-based data sharing fulfills that need.

Editor's note: This article is an extract from "Get the Facts: Real-Time, On-Demand Information for Business Intelligence and Data Integration," available free from the MC Press White Paper Center.

Business moves at the speed of light. At one time, the preceding statement would have been rhetorical. Today, it's often literal.

Consumers shop online, interacting directly with sellers' systems. In the business-to-business world, one company's systems may connect electronically with another company's systems to initiate supply chain activities without any human involvement. And, in some industries, the business environment changes at speeds that were unheard of a couple of decades ago. For example, airlines may automatically adjust fares several times a day to optimize passenger loads and revenues.

In this environment, it isn't enough to make data available sometime, somewhere, somehow. It has to be available right now, wherever it's needed, and in the format that best serves the purposes of its various users.

You've got computers. You've got high-performance databases. Making information instantly available should be easy, right? Wrong.

The problem is that information often resides in applications and databases that typically don't share data well. Consequently, it may require expensive programming and/or regular operator intervention to facilitate data sharing—if it is possible at all.

Specifically, IT is frequently challenged to find easier ways to distribute data between application silos in real-time, without spending a lot of money and without re-hosting applications or performing significant application surgery. Meeting that challenge is difficult because most IT environments include a mix of databases, applications, and operating systems.

There are several possible contributing factors to this data-sharing conundrum. For example, best-of-breed applications running on differing hardware, operating systems, and databases might have been bought for each business function. Mergers and acquisitions might have saddled the organization with a collection of disparate databases. Over time, platform standards might have changed within the organization, but some legacy applications and databases still run on the old platforms. Databases were likely optimized for operational requirements, possibly rendering them extremely inefficient for business intelligence needs.

The upshot is that database administrators now worry about more than simply Oracle, DB2, and SQL Server. Many also have additional databases to support, such as MySQL and CA-IDMS. In addition, there may be hundreds of individual user databases created in MS Access, MS Excel, and Lotus Notes.

The inability to easily share data across disparate platforms and application silos poses a significant obstacle to business strategies that require a more nimble, adaptable, and knowledgeable enterprise. Consequently, senior managers expect IT teams to deliver consistent, quality, real-time data that creates a single view across the business.

Forrester Research refers to "right time data integration," which requires that data assets be constantly recombined for reuse in new applications and new environments. It also means that data integration must occur everywhere in the business, which includes integrating internal systems and online processes with customers and partners.

Software-Based Data Sharing

One solution that can fully meet the needs of right time data integration in environments with disparate hardware, operating system, database, and application platforms is software-based data replication, often also referred to as data-sharing software.

An explanation is in order. The dictionary definition of replication is to make an exact copy, but that is not quite what this class of software does. Data replication products that are used to facilitate disparate data sharing copy the meaning of the data but not necessarily its form.

These products can, for example, share data among DB2, Oracle, SQL Server, Sybase, and other databases. In addition, the data may be organized differently on each of the databases. For instance, for efficiency reasons, an operational database may be structured differently than a data warehouse. Table and column names and data types may also differ in the different databases.

Often, the disparities go beyond differences in database management system (DBMS) technologies and database schemas. Frequently, the data looks different as well. For example, one database might store dates in YYYYMMDD format, while another uses DDMMYY. Product names might be in French in a database supporting operations in France, but in English in the U.S. database. When moving data into a data warehouse, it might be necessary to derive an entirely new "region" field, based on postal/ZIP codes, so that business analysts can quickly and efficiently analyze data on a regional basis. The range of these possible differences is endless.

Data replication products can reconcile these disparities by applying data transformation routines to in-flight data. These products typically ship with a collection of built-in routines to handle standard transformations, such as date reformatting, but also allow you to custom-code your own. The restrictions, if any, on which languages you can use to code transformations vary among the available products.

Software-based replication solutions typically provide graphical interfaces that display the schemas of the databases in the data-sharing processes. You can then point-and-click to map flows between data sources and targets, even when the table layouts and field names differ among the databases.

These graphical interfaces minimize the time required to set up and manage data sharing. In addition, because these tools drive data-sharing processes at the database or journal level, rather than at the application level, they typically don't require any time-consuming programming. The only time additional programming may be required is when your organization needs to perform specialized data transformations that cannot be handled by the routines shipped with the product.

One possible—and often overlooked—use of data-sharing products is to enhance security. Organizations increasingly provide customers and suppliers with access to some segment of their data and applications. The question is, how do you ensure that the data you don't want to share is safe from unauthorized access?

Databases incorporate security features that should protect against these threats, but the better security gets, the better hackers get. Because you can instruct a data replicator to copy only selected rows and columns, one solution to this problem is to maintain a database that contains a near real-time copy of only the data that is to be shared. This database can be located outside the corporate firewall. The firewall can then be configured to allow only the data replicator to send data through the firewall.

Another advantage of this approach is that the structure of the database provided for external users can be different from the structure of the internal database, with each optimized for its specific purposes.

Data Sharing: Top Management Issue

Data-sharing requirements are widespread in today's heterogeneous computing environments. A software-based data-sharing solution can fulfill these needs by providing a simplified, easy-to-manage, click-and-go facility to share data between disparate databases and application silos, without the need for additional programming.

Furthermore, if the data-sharing product is built on open systems technology, its adaptability will extend the life of your existing applications and databases and enable them to remain responsive to changing business and information needs.

For more information, including case studies showing how companies are using software-based data sharing, please download the Vision Solutions white paper "Get the Facts: Real-Time, On-Demand Information for Business Intelligence and Data Integration" from the MC Press White Paper Center.