Presented by:

Kevin Shen

VeloDB

Kevin is currently a Principal Product Manager at VeloDB. Prior to VeloDB, Kevin led various data management products at IBM, such as watsonx.data and IBM Data Virtualization (Watson Query). Aside from working in data management, Kevin has also spent years as a technology consultant at Accenture, working closely with the U.S. federal customers on implementing solutions to reflect legislation

No video of the event yet, sorry!

Most HTAP strategies add complexity rather than removing it. Relying on external pipelines like Flink CDC introduces a heavy middleware layer; you aren't just moving data, you are managing separate clusters, debugging checkpoint failures, and requiring dedicated engineering expertise just to keep the pipeline alive. Conversely, many distributed HTAP databases fail to provide full analytical coverage. They often choke on complex, multi-join queries or suffer from severe performance degradation when analytical and transactional workloads compete for the same resources.

This talk introduces a streamlined architecture built on PostgreSQL and Apache Doris that eliminates these bottlenecks. We will explore Apache Doris’s native CDC integration, which connects directly to PostgreSQL replication slots without intermediate brokers. By removing Flink from the stack, you slash operational overhead and complexity. Attendees will learn how to achieve sub-second data freshness without managing streaming infrastructure, allowing PostgreSQL to focus on ACID transactions while Apache Doris handles complex analytics at scale.

Date:
2026 April 23 11:00 PDT
Duration:
50 min
Room:
Winchester
Conference:
Postgres Conference: 2026
Language:
Track:
Postgres Extensions Day
Difficulty:
Medium