As a leader in the insurance sector, Ping an has prioritized technological innovation, adopting a data-and-business integration strategy to address evolving analytical demands. By continuously upgrading its big data ecosystem, it aims to maximize data value and operational efficiency.
In 2022, the company introduced Apache Doris, an open-source real-time data warehouse, to unify its OLAP stack. This included replacing Presto, eliminating data silos, reducing costs, accelerating delivery cycles, and improving query responsiveness. The shift enabled a move from "extensive" to "refined" growth.
Digital Transformation Journey
- 2005: Established an Oracle-based offline data warehouse and theme-specific data marts.
- 2016: Migrated to Hadoop, later building a data middle platform with governance frameworks.
- 2022: Replaced Presto, HBase, PostgreSQL, and Druid with Apache Doris for unified real-time analytics.
Legacy System Challenges
Ad-Hoc Queries:
- Presto Overload: Unplanned queries forced Presto to process raw Hive data, causing delays and resource waste.
- Governance Gaps: No resource isolation led to chaotic Hive access, excessive temporary tables, and performance bottlenecks.
Multi-Component Complexity:
- Presto, PostgreSQL, Hive, and others created redundant workflows and inconsistent data.
Apache Doris: Unified Analytics Solution
Key Actions:
- Replaced Presto as the primary ad-hoc query engine, eliminating dependency on precomputed metrics.
- Unified OLAP workflows (reporting, multidimensional analysis, customer segmentation) under Doris.
Results:
- Real-time raw data analysis without Presto’s limitations.
- Simplified architecture cut maintenance and development costs by 40%.
- Sub-second query latency for critical business scenarios.
Future Roadmap
- Presto Replacement Extended: Doris will fully replace Presto for data lake queries, enabling seamless lakehouse analytics.
- Multi-Tenancy: Resource isolation to prevent workload interference.
By contributing to Apache Doris’s community, the company reinforces its commitment to open-source-driven innovation.