Why Apache InLong’s Graduation Marks a New Era for Big Data Integration

Apache InLong, a one‑stop massive data integration framework originally contributed by Tencent, has graduated from the Apache Incubator to become an Apache top‑level project.

InLong (named after the mythological dragon “Yinglong”) aims to provide high‑performance, high‑reliability data ingestion and processing for petabyte‑scale streams, supporting automatic, secure, and reliable data transmission as well as stream‑based analytics, modeling, and applications.

Built on Tencent’s internal TDBank and evolved from the TubeMQ message‑queue service, InLong integrates data collection, aggregation, storage, sorting, and ETL into a SaaS platform where users can publish and subscribe to data topics, handling up to 1 quadrillion‑level data streams with high performance and up to 100 billion‑level streams with high reliability.

The platform supports multiple data access methods, including various message‑queue integrations, real‑time extraction‑transform‑load, rule‑based sorting, and provides unified monitoring, alerting, and fine‑grained metrics for visualizing queue status.

Since its incubation in November 2019, InLong has released twelve versions, adding features such as a restructured metadata manager, Flink‑SQL‑based sort‑ETL, and tag‑based cross‑region multi‑cluster support, and is now widely used in advertising, payment, social, gaming, and AI domains.

Its graduation highlights Tencent’s continued contribution to open‑source big‑data ecosystems, following the earlier success of the Angel project, and reinforces Tencent’s position as a leading open‑source contributor.