The first amendment to the United States Constitution requires (among other things) that two powerful entities, religion and government, stay out of each other's business. It's called "separation of church and state," and it's a good idea. Along similar lines, there are good reasons to keep database vendors out of the application space. Although integration between applications is typically a good thing, the integration of the database and application vendor doesn't benefit the customer. In the database arena, data-access standards like SQL, ODBC, OLE DB, and even JDBC are well established and have enabled application-database interaction. And these standards work so well that most major database-application vendors fully support all enterprise-level database platforms. The customer doesn't gain by further integration.

But there are plenty of drawbacks. The biggest is lack of vendor support (real or perceived) for competing databases. One advantage that Microsoft and SQL Server now have in the enterprise ERP space is that the SQL Server database is application-agnostic. Microsoft doesn't yet have a stake in the enterprise ERP market, so there's no technical reason for Microsoft to prefer one ERP vendor to another. Customers are free to choose whatever application best suits their needs.

However, that's clearly not the case with Oracle after its recent acquisition of PeopleSoft. After a prolonged takeover bid, Oracle finally acquired PeopleSoft for $26.50 a share, making the total deal worth about $10.3 billion. Of course, Oracle has been in the application business for a long time, so the acquisition of PeopleSoft really hasn't changed Oracle's fundamental business strategy. However, with the PeopleSoft acquisition, Oracle now offers products that run on other database platforms. The PeopleSoft board of directors accepted the Oracle offer in the best interests of their shareholders, but that doesn't mean it was necessarily in the best interests of their customers.

Many existing PeopleSoft customers, especially those who use databases other than Oracle, are understandably concerned about the Oracle buyout. Probably the most concerned are the JD Edwards customers. PeopleSoft bought JD Edwards in 2003, and most of that customer base runs DB2, although plenty of PeopleSoft customers run on SQL Server. In both cases, customers are concerned as much with Oracle's commitment to support PeopleSoft on other database platforms as with its ability to do so.

Customers must also deal with the vendor lock-in that Microsoft is so often accused of. It's hard to imagine future Oracle/PeopleSoft salespeople recommending a non-Oracle database and easy to imagine them pushing new application sales. Likewise, from the support angle, it's easy to imagine support personnel responding to a database or system disaster with the comment, "This would never happen if you used Oracle."

I hope Microsoft stays out of the enterprise ERP business and keeps SQL Server an application-agnostic database. Separation of database and application vendors is good for the customers because they know they'll get the best support and the most flexibility when each vendor supports its own product and works in partnership with the other. Just as the separation of church and state works for the benefit of the citizens, the separation of database and application vendors works for the benefit of businesses and customers.