Database Design Phase 5: Operation
This article follows on from Database Design Phase 4: Testing.
The operation phase takes place when the testing is complete and the database is ready to be rolled out for everyday use. The users of the system begin to operate the system, load data, read reports and so on. Inevitably, problems come to light. The designers need to carefully manage the database's scope at this stage, as users may expect all their desires to be pandered to. Poor database designers may find themselves extending the project well beyond their initial time estimate, and the situation may also become unpleasant if the scope has not been clearly defined and agreed upon. Project owners will feel wronged if their needs are not met, and the database designers will feel overworked and underpaid. Even when scope has been well managed, there will always be new requirements, These then lead to the next stage.
There are numerous strategies for implementing a rollout. The low-key approach often works well, where the relatively low number of users in the early stage make bug fixing easy. Hugely publicized rollouts often end with egg on the stakeholder's faces, as the best testers of all, the users, invariably find unforeseen bugs, which is best done away from the spotlight. Alternatively, rollouts can occur in a distributed manner, where a pilot branch or office is selected, and when the system has proven its stability, it's rolled out to the remaining branches.
The following are the steps in the operation phase:
- Hand over operation of the database to the users.
- Make any final changes based on the problems discovered by users.
