Summary

Most people don’t need Transparent Gateways between systems; instead they will build a Passive Gateway between systems. It is a minority of Oracle shops that need a Transparent Gateway to turn their many Oracle and non-Oracle databases into one large distributed database.

One reason for this is price. If the Transparent Gateway were free, believe me, the majority of Oracle shops would use it; but a Transparent Gateway is instead more expensive. The Transparent Gateway is a high-end item that is very profitable for Oracle to maintain even if only a small portion of Oracle shops purchase it.

Another reason why the Passive Gateway is more popular is simply that the movement to downsize from the mainframe to an open system is more popular than the movement to hook up existing open systems to mainframes. In most downsizing, we don’t strive to create one virtual distributed database with a mainframe; we instead seek to move applications off of the mainframe and eventually pull the plug on the old steel itself.

The issue of downsizing does fuel the need for coexistence between Oracle and mainframes when we realize that downsizing a system in phases leaves us with dependencies between parts of an IS Operation that have been migrated to an open system and those parts that await migration and still run on the mainframe. When these interdependencies exist, there is a need to share data between Oracle and a legacy system that, in turn, implies some sort of explicit (Transparent, Procedural) or implicit gateway (Passive) to bridge the two systems.

It’s not wise to plan a brand-new distributed database with anything but one database software product running on separate nodes that all run the same operating system. Altering databases and/or operating systems simply slows down the distributed database. This slowdown occurs at all phases from design and planning to actual transaction processing. When we consider mixing mainframes and open systems, this slowdown can be due to an added layer of physical, software, and planning complexity needed to link the two systems.

In reality, most downsizings are planned to minimize dependency between open systems and mainframe components of an IS effort. The planning is such because a greater dependency between systems implies higher software costs. In effect, passive low- maintenance gateways are more popular than Transparent Gateways, replication, and distributed architecture and will remain so until the price of these glamorous and admittedly complex products goes down.

If Oracle gave away its gateway and replication products for free (or almost free), they would lose a great deal of revenue in the short term but would encourage more mainframe dependence with the Oracle database and open systems. It remains to be seen whether Oracle will take this strategic path in the future.