Is SQL Server Ready for Prime Time?

On September 17, 1996, at the DCI Data Warehousing Conference in Phoenix, Arizona, Microsoft announced an alliance with eight other software companies to provide a strong, scaleable, interoperable, data-warehousing solution for SQL Server and Windows NT. The alliance aims to provide users with the ability to create a multivendor, modular, data-warehousing system that connects from one vendor's product to the next easily and uses the best of the NT products in each area.

I met with Jim Ewel, product manager of SQL Server for Microsoft; Gareth Taube, vice president of marketing for Praxis; and Paul Albright, vice president of marketing for Informatica. They explained the blueprint for Microsoft's Active Data Warehousing Framework and the building blocks that will provide its strength and stability.

Infinite Variety
Data warehousing can be difficult to implement. No one product provides all the functionality a company needs in these days of heterogeneous networks, mainframe systems, and legacy databases. Although the process of data warehousing isn't difficult to understand, no two implementations will be exactly alike. You must perform a myriad of functions based on your particular setup. You need a metadata model and the ability to acquire, transform, cleanse, distribute, and replicate data. Additionally, you need the ability to use data in applications and to administer, query, analyze, and report it.

To implement a typical data warehouse, you have to choose various packages that perform the functions you need for the particular configuration of your systems. Then you must find ways for the packages to communicate. In addition, your systems administrator will have to manage multiple databases, data warehouses, data marts, desktop analysis tools, and metadata repositories. The process only sounds confusing because it is.

Microsoft seeks to simplify this process by providing a modular data-warehousing solution with the Active Data Warehousing Framework and SQL Server. The framework will feature a component object model (COM)-based architecture that lets you plug in modules from other vendors to perform the functions you need. The solution will include one metadata repository and one set of administration tasks. The framework will enable alliance-compliant third-party vendor packages to link and work together in a simple and efficient manner.

For example, if your enterprise has DB2, Oracle, and SQL Server databases and you don't want to eliminate any of them, you can keep them all in synch, enable individual data marts to subscribe only to the data they need, and make implementing and administering your data warehouse easier and faster at the same time. Using the Active Data Warehousing Framework, products from Praxis International (heterogeneous data replication to keep your various databases in synch), Platinum Technology (individual data marts by subscription), Informatica (data transformation), and other alliance-compliant vendors will be able to interface with each other and with SQL Server to provide these functions. Table 1 lists the basic capabilities and expertise that the alliance members bring to the framework.

Quiet, Please!
This limited and incomplete example doesn't begin to show the breadth and depth of solutions this framework supports. You'll be able to implement your data warehousing solution in a heterogeneous environment with different legacy databases without the expensive and complex programming effort currently required--and in a fraction of the time. If the Microsoft Alliance for Data Warehousing and its Active Data Warehousing Framework live up to their potential, the results should once and for all quiet those pundits who claim that SQL Server isn't ready for enterprise prime time.

--Jane Morrill

End of Article