User Story: How Tencent Music saved 80% in costs by migrating from Elasticsearch to Apache Doris

[httpshirley]

Sharing another Elasticsearch-to-Doris migration story with the community!

Use case: Tencent Music Entertainment (NYSE: TME) needs to unify their full-text search, audience segmentation, and aggregation analysis workloads into one platform, to get rid of the high costs and bottlenecks brought by their old hybrid architecture.

Why the change❓
Elasticsearch was powerful for search, but it lacked support for complex queries like JOINs and came with a large storage footprint. Apache Doris, with its inverted index, not only supports full-text search but also delivers high-performance analytics with better storage efficiency.

What they gain 🎁
After the migration, TME benefits from a cost reduction of 80%, 4x faster data ingestion, and the convenience of a unified data platform.

What they share (tips and tools) 💡 🛠️
As an experienced user of Apache Doris, TME generously shared valuable insights into their use of Doris' inverted index, strategies for resource isolation across business lines, and best practices for business migration. In addition, they have developed and open-sourced their own platform, SuperSonic, which enables seamless data migration, unified data governance, and natural language-driven data analytics — available for interested users to explore and reference.

Full post: https://doris.apache.org/blog/tencent-music-migrate-elasticsearch-to-doris