Analysis of Current Trends, Architecture, and Technologies in Chinese Database Products

The content references the "2022 China Database Product Strategy Analysis Report" and the "2021 China Distributed Database Market Report" as primary sources.

It analyzes future development trends in database management, emphasizing technology borrowing, integration, and continuous evolution to meet market demands.

Database development stages are defined, describing databases as collections of data stored on computer devices and organized according to specific models for shared access.

A Database Management System (DBMS) is explained as software that provides services such as data definition, storage, backup, access, update, analysis, security, and operational management, with data models forming the core foundation.

Database design processes aim to provide more effective semantic relationship expression and offer semi‑automatic or automatic design tools at each stage.

Designing an optimal database schema for a given application environment involves constructing the schema, building the database and its applications, and ensuring efficient data storage and classification.

Chinese database vendors are categorized into traditional, emerging, cloud, and ICT‑cross‑industry providers, each offering various centralized and distributed database products.

The overall database architecture consists of management, compute, and storage modules, with a physical resource layer providing the underlying infrastructure.

The physical resource layer supplies the basic environment for databases and upper‑level business systems.

The compute module parses query requests, generates execution plans, and distributes them across compute nodes for parallel execution.

The storage module carries out data operations from the compute layer and ensures persistent storage, while the management module coordinates distributed clocks, maintains metadata, and provides configuration and monitoring interfaces.

HTAP (Hybrid Transactional/Analytical Processing) currently follows two architectures—separated and unified—with the separated architecture being dominant; cloud‑native environments are driving new HTAP solutions.

HTAP eliminates the gap between OLTP and OLAP, enabling a single distributed database to serve both transactional and analytical workloads for real‑time decision making.

Key challenges include minimizing resource interference, ensuring high data visibility, and achieving low latency when co‑running OLTP and OLAP workloads.

Typical HTAP application scenarios are data‑intensive businesses (e.g., IoT, healthcare, risk control, personalized recommendation) that embed analytics into transaction systems, and real‑time data service platforms that treat data as a consumable asset.

Compression storage technologies such as Zstd (high compression ratio and efficiency) and Iz4 (fastest compression/decompression for OLAP) are discussed, highlighting their suitability for cold storage and analytical workloads.

Data compression aims to reduce transmission and storage costs while preserving information, improving efficiency, and eliminating redundancy.

Common compression algorithms include Snappy, Terark, zlib, bz2, lz4, lz77, zstd, brotli, B‑tree, RLE, Delta Value Encoding, Deflate, and dictionary‑based methods.

Storage index structures such as hash tables, B/B+/B* trees, LSM trees, R‑trees, inverted indexes, matrix storage, object/block storage, and graph storage determine the performance and capabilities of storage engines.

Modern databases typically employ Hash, B+Tree, or LSM‑Tree storage engines for their primary index architectures.

Intelligent database governance is essential for security and scalability; traditional rule‑based tools are insufficient, prompting the adoption of AI‑driven optimizers such as Al4DB.

Database governance now requires cloud‑based automation, AI‑driven parameter tuning, data‑driven self‑monitoring, and intelligent self‑diagnosis to reduce reliance on DBAs.

Intelligent parameter tuning combines deep reinforcement learning and global search algorithms to automatically obtain optimal configurations; vendors offering this include Huawei Cloud, Tencent Cloud, OceanBase, Baidu Cloud, and others.

Deployment models are divided into traditional on‑premise installations, which depend on high‑end hardware and lack scalability, and cloud deployments that enable massive horizontal scaling across servers and virtual machines.

Cloud‑deployed database products have entered a mature commercial stage, offering hosted, service‑based, and cloud‑native forms.

The article concludes with a reprint notice and promotional links encouraging readers to purchase a compiled collection of architecture‑related technical materials.