OceanBase 2.0 Technical Overview: Enabling Ant Group’s Double 11 Mega‑Promotion

Introduction

The 2018 Double 11 shopping festival set a record of 213.5 billion CNY in transactions, and the Ant Group’s payment system faced unprecedented load. OceanBase, now in its fifth year of supporting the event, powered the core payment link with OceanBase 2.0, delivering a 50 % performance improvement and zero‑cost scaling.

1. Core Technologies Used in the 2018 Double 11 Promotion

OceanBase 2.0 was deployed on the core payment chain, providing an elastic distributed database capable of handling millions of payments per second.

The entire core link ran 100 % in containers, and a storage‑compute separation architecture reduced resource costs while strengthening a unified storage foundation.

OCP (OceanBase Cloud Platform) introduced AI‑driven SQL optimization, fault root‑cause analysis, and intelligent capacity planning, merging expert knowledge with machine‑learning algorithms.

Orchestration capabilities allowed rapid rollout and rollback of changes; over 40,000 automated changes were executed during the promotion with zero incidents.

These intelligent operations and platform abstractions were also packaged for external commercial scenarios.

2. OceanBase 2.0 & Million‑Payment Architecture

To meet the growing demand for a million‑transactions‑per‑second capability, OceanBase 2.0 introduced fine‑grained sharding that can split data across an unlimited number of machines, delivering extreme elasticity without impacting applications.

The architecture avoids invasive changes to business logic, using partition groups for automatic routing and aggregation, eliminating distributed‑transaction overhead and achieving linear, lossless scaling. Multi‑version, share‑nothing design ensures high availability and fault isolation.

Compared with OceanBase 1.0, version 2.0 upgraded native sharding, distributed‑transaction handling, and commit‑log overhead, resulting in a 50 % overall performance boost and no additional hardware for the Double 11 event.

3. Containerization & Storage‑Compute Separation

Resource cost during peak periods is minimized by rapidly allocating and releasing containers. OceanBase’s containerization abstracts underlying resource differences, enabling fast scheduling across offline, cloud, or other clusters.

Storage‑compute separation deploys compute and storage on distinct substrates, allowing independent scaling. During peak traffic, compute resources can be expanded without proportionally increasing storage, while in normal operation storage I/O is not a bottleneck, improving overall utilization.

Typical scenarios include:

Peak‑time CPU surge with minimal storage growth – compute nodes are scaled up, reducing cost and improving elasticity.

Daily operation where LSM architecture turns random I/O into sequential I/O – storage‑compute separation reduces fragmentation and enables shared storage pools.

4. Platform Intelligence

OCP combines operational expertise with AI/ML to create intelligent models for SQL plan validation, fault diagnosis, and capacity forecasting. Under massive load, it performs minute‑level plan correctness checks and auto‑remediates risky SQL.

Capacity modeling predicts resource needs per tenant, cluster, or Docker instance, enabling automatic scaling and optimal load balancing, achieving near‑full resource utilization (≈(n‑1)/n).

OCP acts as the “intelligent brain” of OceanBase, monitoring everything from individual SQL plans to thousands of machines, paving the way for autonomous database operations.

5. Ecosystem & Connectivity

Beyond internal use, OceanBase is being opened to external industries, leveraging its technology to accelerate digital transformation across enterprises. The platform’s ecosystem now includes ISV collaborations, extended product offerings, and a focus on high‑availability, high‑performance, low‑cost database services.