How Cloud‑Native Databases Are Redefining the Future of Data Management

Cloud Database Market Shifts

The cloud wave is reshaping the traditional database market, with cloud databases gaining roughly 10% market share and growing rapidly, challenging the dominance of legacy vendors like Oracle, Microsoft, IBM, SAP, and Teradata.

According to Gartner, the global database market is about $40 billion, with China accounting for 3.7% (~$1.4 billion). Traditional relational databases still lead, but newer systems such as MongoDB, Redis, and cloud‑native offerings are diversifying the landscape.

Why Databases Are Critical in the Cloud Era

Databases connect IaaS infrastructure with SaaS applications, making them a pivotal component for AI, computer vision, and other intelligent services.

Four major database categories—OLTP, OLAP, NoSQL, and database‑service tools—drive the focus of cloud database providers.

Evolution of Database Architectures

Historically, databases used a shared‑memory (single‑node) architecture. As scale demands grew, shared‑storage (distributed storage with RDMA) emerged, exemplified by Alibaba Cloud’s POLARDB, offering one‑write‑multiple‑read capability, elastic scaling, and high availability through three‑replica storage and a modified Raft protocol.

Shared‑nothing architectures address the scalability limits of shared‑storage; POLARDB‑X combines shared‑nothing with shared‑storage to achieve horizontal scaling and cross‑region strong consistency, supporting massive concurrency for financial‑grade workloads.

Core Pillars of Cloud‑Native Databases

Multi‑model support : Both north‑bound (single KV store with diverse query interfaces) and south‑bound (heterogeneous storage types like graph, vector, document) models, e.g., Azure CosmosDB.

Intelligent/autonomous management : Self‑awareness, self‑decision, self‑recovery, and self‑optimization, with machine‑learning‑driven tuning.

Hardware‑software co‑design : Leveraging RDMA, NVM, GPUs, and FPGAs to reshape storage engines.

High availability : Redundant replicas, active‑active design, and CAP‑aware trade‑offs (typically Consistency + Partition tolerance).

POLARDB Highlights

POLARDB uses a shared‑storage architecture with one‑write‑multiple‑read clusters, achieving up to 100 TB storage and 1 million QPS per node. It provides MySQL and PostgreSQL compatibility, active‑active replication, rapid elastic scaling (2‑core to 32‑core in minutes), and continuous recovery via shared physical logs.

Its proxy layer (PolarProxy) handles read/write separation, load balancing, and high‑availability failover.

POLARDB‑X for Extreme Scale‑Out

POLARDB‑X adopts a shared‑nothing design, supporting cross‑AZ strong consistency and handling peak traffic such as Alibaba’s Double‑11 shopping festival, with proven performance in internal production.

AnalyticDB – OLAP Benchmark

AnalyticDB features a hybrid row‑column engine, vectorized processing, and can analyze petabyte‑scale data in seconds. It topped the TPC‑DS benchmark for price‑performance and integrates with DTS for real‑time data pipelines.

Autonomous Database Platform (iBTune)

The Self‑Driving Database Platform (SDDP) collects performance metrics and applies machine‑learning models to dynamically adjust buffer sizes, saving up to 20 TB of memory across 7 000 instances.

Security and Cloud‑Native Services

Alibaba Cloud implements data‑at‑rest encryption, BYOK support, and plans for in‑memory encryption and trusted logging.

Comprehensive database services span relational, graph, and analytic engines, with migration tools (DTS) and end‑to‑end operational management.

Overall, cloud‑native databases represent a new battlefield, offering elasticity, scalability, and intelligent automation that push the 40‑year‑old database industry toward unprecedented capabilities.