An upgrade of a large non-Unicode application system to a Unicode-based system is normally associated with high costs in terms of human resource,hardware, and especially with long down time. We have developed a method to upgrade a non-Unicode database to a Unicode based system with no down time, with no additional hardware, and with minimum human resources.

Non-Unicode database tables are copied into Unicode tables one by one. As soon as the Unicode copy of a table is created, it replaces the original, non-Unicode table. During this operation, the non-Unicode system does not need to be shut down. The method requires no additional hardware because the copy is executed on the original database and it creates a duplicate of only one table at a time. The method is suitable also for systems larger than 1 Terabyte. Such system upgrades normally require the creation of a new database. In our experience, the method shortens the total upgrade time by a factor of 5 compared to the conventional upgrade method.

We would like to share the concepts of the method and our experiences with the attendees of the IUC in order to promote the Unicode standard also for large database-based applications.