TCLake delivers integrated management for structured, semi-structured, and unstructured data built on a unified data foundation. By seamlessly converging workloads like batch and stream processing, it drives deep synergy across multimodal data, enabling unified storage and governance for the Data+AI era.
Batch-Stream Unified Table Format
Apache Iceberg Compatibility: Built as a non-intrusive extension of Apache Iceberg, it fully supports batch-stream unified scenarios while maintaining compatibility with native Iceberg semantics and its rich open-source engine ecosystem.
Advanced Near-Real-Time Lakehouse: While native Apache Iceberg restricts downstream streaming consumption of updated data, TCIceberg seamlessly supports streaming writes alongside CDC (Change Data Capture) incremental reads. It also features an extensible merge process to handle advanced workloads like partial column updates and pre-aggregation.
Performance Enhancements: Leverages an auto-bucketing mechanism to significantly boost Merge-on-Read (MoR) performance during data update scenarios.
Intelligent Data Optimization: Features real-time monitoring of table read/write operations. Based on these telemetry insights, TCIceberg automatically schedules optimization resources on-demand, maximizing both the quality and efficiency of data maintenance.
Unified Data Catalog
Multimodal Data Catalog: Features a natively built-in multimodal data catalog service covering tables, unstructured volumes, machine learning models, views, and functions. It delivers comprehensive, full-lifecycle catalog management for all Data+AI assets.
External Asset Integration: Beyond the native catalog, it seamlessly integrates catalogs from heterogeneous data sources and distributed systems (e.g., MySQL, Apache Hive, Apache Doris). This empowers users to access and manage both TCLake and external data assets through a unified view, effectively breaking down data silos.
Unified Access Control: Implements an RBAC-based unified access model across all data catalogs. By providing a standardized access layer, it establishes a robust access governance framework that spans the entire data lifecycle.
Open Engine Ecosystem (Ongoing Integration)
Tencent Cloud Ecosystem: Seamlessly integrates with Tencent Cloud's native analytical suite, including EMR, DLC, and TCHouse. It delivers an out-of-the-box experience, allowing users to instantly leverage the built-in mainstream computing engines of these services.
Open Source Ecosystem: Natively supports major open-source big data computing engines like Apache Spark and Flink, alongside leading AI training frameworks such as Ray and TensorFlow.
Serverless & Zero-Ops
Fully Managed Services: Delivers fully managed, out-of-the-box data catalog and storage services, completely freeing users from the burden of provisioning and maintaining complex underlying infrastructure.
Intelligent Data Management: Automatically executes background optimization tasks—such as small file compaction, stale snapshot cleanup, and comprehensive data lifecycle management—requiring zero manual intervention.