Friends (CQRS)

This story demonstrates the CQRS (Command Query Responsibility Segregation) pattern: how Actionbase became a read-optimized view layer for KakaoTalk’s friend relationships.

The Challenge

Friend relationships in KakaoTalk are stored across dozens of sharded MySQL databases. To serve read-heavy queries, an HBase-based view layer was already in place:

flowchart LR
    App[Application] -->|write| MySQL[(MySQL Shards)]
    MySQL -->|sync| HBase[(HBase View)]
    App -->|read| MySQL
    App -->|read| HBase

This worked for the original use case. But as requirements evolved, limitations emerged:

• The HBase view was designed for a specific access pattern
• Limited flexibility for new query types and schema changes

Actionbase got the chance to prove itself here—not as a replacement, but as an additional view layer.

Integration Strategy

Rather than replacing the existing system, we added Actionbase alongside it:

flowchart LR
    Existing[Existing System] --> Debezium[Debezium CDC]
    Debezium --> Kafka[Kafka]
    Kafka --> AB[(Actionbase)]
    App[Application] -->|write| Existing
    App -->|read| Existing
    App -->|read| AB

The key insight: Actionbase doesn’t need to be the source of truth. It can serve as a flexible view, consuming changes via CDC. The existing system (MySQL + HBase) remained unchanged.

Stage 1: CDC Pipeline

First, we set up the data flow. Debezium captured changes from MySQL shards and published them to Kafka. Actionbase consumed these events and applied mutations.

Stage 2: Bulk Load

For historical data, we:

1. Dumped the MySQL tables
2. Bulk-loaded into Actionbase
3. Replayed the WAL to catch up with changes during the dump

Note: The migration pipeline (bulk loading) is currently internal. Open source release is in progress — see Roadmap.

This gave us a consistent snapshot without downtime.

Stage 3: New Query Layer

With Actionbase in place, new systems used it for all query types—get, scan, count, and reverse—through a single schema definition.

What We Learned

• Actionbase works as a CQRS view. It doesn’t have to own the data to add value.
• CDC enables non-invasive integration. No changes to the write path required.
• Schema flexibility unlocks new use cases. Get, scan, count, reverse—all from a single schema definition.

This pattern opened the door to adding Actionbase alongside existing systems without migration risk.