The Challenge
SAS data pipelines needed to be migrated to GCP before license expiration
- Outdated technology & unwanted costs: Legacy outdated pipelines written in SAS code needed to be migrated to the Cloud in order to unify all data management task in a single platform and save unnecessary costs.
- Orchestration design need: Orchestration based on time of execution, lack of documentation for data dependencies and lack of expertise of the technical team on cloud orchestration.
- Data source structure not identical: Data model and sources used in SAS platform were different than the ones in GCP.
- Uncertainty on business relevance: There were no clarity on business relevance for all data pipeline outcomes.
Our Solution
Leverage GenAI for Automated SAS Code translation/migration
- We developed a Gen AI automated tool to extract from SAS code all data sources and outputs in order to create a data dependencies chart and be able to orchestrate properly all data processes.
- The Gen AI tool also assisted in translated the SAS Base code to the appropriate cloud-based programming language, adapting the code to fit the new data model and optimizing cost & performance effectiveness.
- Manual assessment of business relevance was needed to refine which data processes needed to be actually migrated.
KEY Benefits
Process migration in 4 months
Unnecessary costs removed
Process migration completed in 4 months, meeting a tight deadline
$ 200K/year SAS License
$ 150K/year Opex Savings
RESULTS
- Performance improvement of +30%
- Unification of all data pipelines for this team in a single platform, improving governance and simplifying maintenance.