Technology Conglomerate Seeks to Migrate from Teradata
A multinational technology conglomerate needed assistance with its database services after encountering high licensing costs and performance problems. The conglomerate manufactures and sells networking hardware, software, and telecommunications equipment alongside other high-tech services and products.
Business Challenges Surrounding Teradata and Database Performance
The company wanted to migrate away from its legacy solution to a new modern data warehousing platform. This would involve migrating away from an existing Teradata database platform to a cloud-native Snowflake data warehouse.
Teradata used a specific SQL syntax within SAP Business Objects, and the company was in the process of migrating to Snowflake. These SQL syntaxes needed to be rewritten and reintegrated to work in the new Snowflake platform.
A Proof-of-Concept Design for Teradata-to-Snowflake Conversion
Trianz designed a Proof of Concept to automate the conversion of SQL queries. This included five web intelligence reports contained within one centralized business objects universe. Then, a semantic layer conversion tool was used to automate SQL conversion between Teradata and Snowflake.
The timeline from project commencement to completion was set at one month.
How Trianz Approached Teradata-to-Snowflake Migration
A team from Trianz, consisting of two Snowflake data engineers and one Business Objects architect, oversaw this project. They started by analyzing and documenting SAP Business Object requirements, designs, and business processes. This would assist with successful implementation during migration to Snowflake.
Next, the Trianz EVOVE platform was used to translate SQL syntaxes into a Snowflake compatible format, before Trianz reintegrated converted syntaxes in a centralized Business Objects Universe.
Around 90% of the SQL tables were automatically converted, with the final 10% requiring manual intervention by Trianz. Considering the technology client was manually converting queries, the switch saved them an immense amount of time.
To validate the conversion, Webl reports were tested against the new Snowflake environment. This would ensure that reports run successfully, and that data matches in both Teradata and Snowflake. All reports were executed successfully alongside data matching checks.
Transformational Effects After Migrating to Snowflake
Around 75% of Business Object instances were programmatically converted to Snowflake during this project. Trianz EVOVE was a great help here, automating the conversion of more than 95% of derived tables, with six tables requiring manual conversion.
Within the first week of project commencement, 40% of reports had achieved the intended result. Within a month, the entire conversion and migration process was finalized. Typically, such a process would take up to three months to complete manually.
Additionally, Trianz identified seven tables outside of the project scope that could be optimized for use in Snowflake. The client could also scale up in future with five or six SAP universes, future-proofing its data operations on Snowflake.
| Universe Business Objects | Count |
|---|---|
| Converted Programmatically | 460 |
| Converted Manually | 219 |
| No Conversion Needed | ~300 |
| Total | ~1900 |
| Derived Tables | Count |
|---|---|
| Converted Programmatically | 100 |
| Converted Manually | 6 |
| Total | 106 |
Break-down of Programmatic Errors
| Resolved after compiler fix | Manually fixed items | Underlying database items | No issue items |
|---|---|---|---|
245 | 219 | 35 | 30 |
