Case Study: InSQL to PI Migration
- InSQL 1:
- Time span: 4 year and 7 months.
- No. of Tags: approx. 23 000.
- InSQL 2:
- Time span: 1 years and 11 months.
- No. of Tags: approx. 40 000.
|The result of migrating historical data from InSQL 1 and InSQL 2 to PI|
|Migration Analysis: Source (InSQL) Plot|
|Migration Analysis: Target (PI) Plot|
- Synchronisation of tag configurations: This consists of linking tags from the source historians (InSQL 1 and InSQL 2) to the corresponding tags in the target historian (PI). This step is particularly important for tags whose names differ in the source and target historians.
- Create the Migration Jobs: Migration Jobs are responsible for the actual migration and run as services. A Migration Job is characterised by source and target historians, a tag list (imported from Tag Manager), and start and end times. Migrations Jobs are subdivided into batches, which are the smallest repeatable units of a migration job. The migration was completed using 6 jobs (4 for InSQL 1 and 2 for InSQL 2) and each batch had a time range of one day.
- Analysis and verification: After a batch completes data migration, there is an option of running an analysis step. During this step, data from the source and the target are extracted and compared. Based on the difference between the source and target data, thresholds can be set to identify potentially problematic cases. Because of this, the migration was not merely a blind process. The results of the migration were verified.
- Tag configuration synchronisation and importing to Tag Manager (InSQL 1 and InSQL 2): 2 weeks.
- Executing the migration jobs (including analysis):
- InSQL 1 (23 000 tags, 4 years and 7 months): 2 weeks.
- InSQL 2 (40 000 tags, 1 year and 11 months): 5 days.