H1

H2

H3

H4

H5

H6

Paragraph

BOLD

ITALIC

LINK

BLOCK QUOTE

BLOCK quote

Preformatted

Metric	Impact
User Adoption	3x increase in user adoption — from 400 to 1,500+ users
Pipeline Reduction	700+ duplicate pipelines eliminated
Cross-Platform Usage	6x increase in cross-platform queries
Query Volume	2x increase in query volume
Performance	3x faster query performance compared to the previous setup
AI Enablement	300+ AI/ML models supported via Starburst
Self-Service Impact	Starburst was adopted across 230 departments, 65 business units, and seven geographies
Hue Dependency	70% reduction in Hue usage for data exploration
Job Orchestration	Thousands of daily jobs are now running through the Starburst-powered pipeline framework

1. one
2. two
3. three

With over 6,000 TB of data across 300+ systems and 1M+ attributes, OCBC was looking to modernize its analytics infrastructure, unify access to distributed data, and enable real-time, AI-driven insights. Starburst helped OCBC transform into a next-gen, AI-ready enterprise by virtualizing access across platforms and delivering federated analytics at scale.

 

 

 

Challenge

OCBC’s data infrastructure spanned two major platforms

• Teradata – a long-established enterprise data warehouse for batch-processed, logic-rich historical data
• Cloudera Hadoop – a newer big data platform for raw, unstructured, and high-volume workloads

Despite the strength of each platform individually, the lack of seamless integration created multiple challenges:

• Duplicate data pipelines for each system
• Tedious data movement processes to unify datasets
• Performance bottlenecks affecting SLA commitments
• Limited ability for business users to self-serve or run cross-platform queries
• Growing demand from 1,400+ analytics users across 230 departments and 65 business units

The fragmented environment hindered scalability, slowed AI model execution, and restricted the bank’s ability to democratize data access.

Solution

To overcome these limitations, OCBC adopted Starburst Enterprise as a federated query engine—creating a unified access layer across Teradata and Cloudera without moving data.

• Key components of the solution:
• Cross-platform query virtualization: Users query both platforms from a single entry point
• In-memory compute & pushdown optimization: Preserves SLA-sensitive workloads in both Teradata and Hadoop
• Self-service data pipeline framework: Built using Starburst + dbt + Dagster, empowering business and technical users alike
• AI model enablement: All ML/AI pipelines and campaign analytics are now powered by Starburst