Big Data 15 min read

Douyin Group Data Asset Management Platform and Data Lineage Architecture Overview

This article provides a comprehensive overview of Douyin Group's data asset management platform, detailing the evolution, architecture, and applications of its large‑scale data lineage system, and discusses future directions for enhancing data quality, cost efficiency, and security across the organization.

DataFunTalk
DataFunTalk
DataFunTalk
Douyin Group Data Asset Management Platform and Data Lineage Architecture Overview

The article introduces Douyin Group's one‑stop data asset portal, emphasizing the shift from traditional metadata to a systematic "manage, find, use" data asset platform that supports diverse data sources and provides unified metadata lakes with full‑link lineage.

The platform enables asset evaluation, search, recommendation, and AI‑driven discovery, facilitating data consumption through portals and productized capabilities.

Four key topics are covered: overall data lineage overview, system architecture, application scenarios, and future outlook.

1. Data Lineage Overview – Douyin aims to build comprehensive, real‑time, accurate big‑data lineage to empower various scenarios, recognizing lineage as the core of metadata.

The lineage covers source‑level, production‑level (real‑time and offline warehouses), and application‑level relationships, with metrics such as coverage, accuracy, and completeness.

2. System Architecture – Challenges include fine‑grained parsing, unstructured sources, cross‑region coverage, and large‑scale application tracing. The solution comprises data source ingestion, metadata and lineage collection, storage using graph databases (JanusGraph, Neo4j, NebulaGraph), and analysis services.

Parsing leverages Antlr and Calcite to handle multiple dialects and complex scripts, converting syntax trees into lineage information.

Three lineage services are described: production lineage, cross‑region lineage, and application lineage, each addressing specific data flow and tracing needs.

3. Application Scenarios – Includes data development (impact assessment, field lineage, task acceleration), data governance (risk asset identification, cost calculation, timeliness and accuracy assurance, security), and data asset utilization.

Use cases demonstrate how lineage supports cost reduction, quality improvement, and efficient data operations.

4. Future Outlook – Plans focus on full coverage, standardized lineage APIs, open contributions, finer granularity (row‑level lineage), and deeper exploitation of lineage for quality, efficiency, and security.

Douyin's DataLeap platform aims to continuously enhance metadata, integrate active metadata and large‑model capabilities, and expand to the broader market.

Big Dataplatform architecturedata lineagedata governancemetadata management
DataFunTalk
Written by

DataFunTalk

Dedicated to sharing and discussing big data and AI technology applications, aiming to empower a million data scientists. Regularly hosts live tech talks and curates articles on big data, recommendation/search algorithms, advertising algorithms, NLP, intelligent risk control, autonomous driving, and machine learning/deep learning.

0 followers
Reader feedback

How this landed with the community

login Sign in to like

Rate this article

Was this worth your time?

Sign in to rate
Discussion

0 Comments

Thoughtful readers leave field notes, pushback, and hard-won operational detail here.