Identifying data constrained activities for migration planning

Abstract

Constant and rapid changes in the market place have inevitably brought changes to business process including the long running process. When changes are mandatory or require retroactive compliance, it may be necessary to migrate existing running process instances to a new version model. Current dynamic instance adaptation approach requires each instance to adapt to the changes individually. To support multiple instance migration across process models, we present an approach where we leverage model comparison technique; control and data flow analyses to identify data constraints for changed activities and identify potential migration points where we could migrate the instances. We then capture the migration point information in a migration plan which can later be used in the runtime environment to orchestrate the migration. We have implemented a prototype of our approach to demonstrate its applicability. © 2009 IEEE.