The Swiss Federal Railways (SBB) transport every year some 250 million passengers, and up to 45 million tons of freight. Each day, all its trains combined cover a distance equal to more than 8 times the circumference of the globe. SBB-trains are calling at some 800 stations throughout the country. Just under 33,000 personnel master their task with the help of a highly developed technical apparatus. The yearly turnover exceeds 6.5 billion francs.
SBB started using CA Gen in 1992 with IEF 5.0. SBB has 125 toolset licences and use three MVS/DB2 host encyclopaedias. These contain over 300 models, divided into 70 discrete projects. There are a total of 24 million objects in a 16Gb database. 30 projects are now in production.
Every project has its own models (design, test, integration and production). The shared data objects (corporate entity types) and the common design services (action blocks for security, string conversion, help, navigation systems) are taken from the models where they have been developed, then saved in one of the shared object models before being migrated to the other models that need them.
GuardIEn was primarily purchased in 1996 to manage migrations and installations to the production environment. GuardIEn change requests are created to define and control the changes to the models. With the GuardIEn report utilities and some extension reports from SBB the model co-ordinators look for the changes in their models and select the deliverables which have to be installed in the next environment. The next steps are automatically assisted with the tool.
In the second part of 1997, when all projects will be managed with GuardIEn, SBB want to expand the use of the tool to all developers in large projects for migration and installation of deliverables in the test environment as well as the production environment.
Before the introduction of GuardIEn, the processes used to update production were not precisely defined and controlled. Because many of the tasks were not automated, each project had to interpret how they should perform their part of the process. This lack of standardisation lead to a lack of adequate status information, a lack of history of changes applied to production and consequent difficulties in managing the production environment.
The introduction of GuardIEn forced an analysis of the production update processes, requiring us to describe them with a better precision. We now have the assurance that the processes are run correctly with a standard approach to documentation of project history.
CA Gen only reports on one environment (model) at a time. With GuardIEn we have the capability of tracing the changes across all models in the projects and introducing explicit version control on the project’s deliverables.
The main benefits realised through the use of GuardIEn are:
- reduction of the migration effort.
- reduction of timestamp conflicts between modules and bind members
- reduction of the installation effort
Before the introduction of GuardIEn, the model co-ordinators performed the production updates by using a combination of CA Gen change reports, aggregate sets, intelligent regeneration and a very detailed knowledge of the encyclopaedia. Furthermore, to prevent timestamp conflicts, most projects decided to re-generate the complete business system even if there had only been a few changes.
Using GuardIEn we can identify the precise changes, scope them into a change request and the GuardIEn production update ensures that the changes are correctly implemented. This has resulted in a large reduction in effort for both the project teams and the central CA Gen support staff.
We have been able to expand our level of control from the model level to an overall project level that comprises all of the models and detailed version control at the object level.