By David Greer, MB Foster Associates
Many customers decide to stay with the HP 3000 because of its extraordinary reliability and the low cost of ownership. When working with both homesteading and migrating customers, we see a number of practices around application and data management that can provide benefits now and in the future.
HP 3000 applications have been developed over many years. This makes the applications highly effective to organizations because they accurately reflect the business rules of the organization. Any application built over time faces challenges matching all source code to all running production code. Many HP 3000 sites do not have a formal change management process for their applications.
Change management typically is implemented in two parts: version control and governance. Did you know that you can put your HP 3000 source code under the control of a version control system such as Microsoft Visual Source Safe? Doing so allows an organization to identify and document all component pieces of each application. The effort and knowledge gained reduces the risk to the organization, by formalizing the knowledge that is often scattered around many individuals.
A governance process for the release of new versions of HP 3000 applications further reduces the risk of changes. A version control system helps. It causes organizations to assign version numbers and identify all specific files that need to be changed to implement an application change. A formalized development, test and release governance process makes sure that IT, users, and management are aligned when it comes to releasing new versions of the software. Not only does this reduce organizational risk on numerous fronts, it sets up an organization for future change. We have yet to see a successful migration that did not have strong change management and governance.
A second area where HP 3000 sites can improve performance is in data management. Redundant data can cost organizations millions of dollars every year. As many HP 3000 databases have been developed over decades, they often have large amounts of duplicate data. We have observed and participated in cases where rationalizing both duplicate data and the amount of historic data has resulted in large space reductions while speeding up batch processing by over ten times.
Another major focus area for HP 3000 improvement is cleaning up data. We advocate that you set up simple job streams to check for bad data nightly or weekly and report it directly to affected users. At a recent conference, the Registrar of a major Canadian college reported they now check all data nightly and mail suspected bad data directly to users responsible for entering it.
Removing bad data reduces the amount of data in your database and insures that people who depend on this data make the right decisions. In our migration work, it is common to have to spend a lot of time cleansing data before migrating it — much quicker if your data’s already clean.
We have seen many HP 3000 sites leverage a data mart. A data mart provides an alternative view of your HP 3000 data in a popular SQL database such as SQL Server. While introducing redundant data, the benefits outweigh the costs, especially when the data is transformed as part of the replication process. Using SQL Server allows HP 3000 sites to hire and train experts in the latest technologies speeding delivery and lowering costs.
Some sites are building all new functionality on top of MS SQL Server using bidirectional database replication technology. Over time the business becomes less dependent on the HP 3000 application. It also ensures that if you do migrate, the majority of your interface points are to SQL Server.
Many of these ideas can be introduced as a project or one application at a time, letting you spread out the implementation cost. The biggest hurdle is making the commitment to change the way you do things now, to increase your ability to execute in the future.