Industrial Bank

Customer Overview

With commercial banking business as the core, its total assets exceed 10 trillion yuan. It shows strong strength in the financial market, investment banking, and other fields, and its market influence has significantly improved.

Industrial Bank announced a comprehensive upgrade to its Bank-Bank Platform and launched the Internet wealth management brand "Qian Da Zhang Gui", promoting the cooperation and development of multiple businesses. However, with the rapid growth of the business, the existing Informix database cluster faces performance bottlenecks, the database version is old, the cost of hardware upgrades is high, and the reserve of technical talents is insufficient.

Solution Design

Requirements Analysis

Improve system performance: Solve the performance bottleneck of the existing Informix database cluster and improve the business processing capability and response speed.
Reduce operational costs: Reduce the high costs incurred by hardware upgrades and maintenance of the old-version database.
Enhance technical support capabilities: Choose a database product with extensive community support and a mature ecosystem to enhance the subsequent technical support and maintenance capabilities.
Adapt to business development: Adopt a more flexible and scalable architecture to support the rapid development of future businesses.

Technology Selection

Architecture selection:

Distributed microservices architecture: Replace the existing centralized architecture to provide higher scalability, flexibility, and fault tolerance.

Database selection:

MySQL: As a representative of open-source databases, MySQL has extensive community support, a rich ecosystem, and good performance. Choosing MySQL as a replacement for Informix database can reduce migration costs and enhance technical support capabilities.

Implementation Plan

DMP Cluster Deployment

DMP platform installation and configuration: Install and configure the Action Technology DMP platform according to the needs of Industrial Bank to ensure the stable operation of the platform.
MySQL business database migration and splitting: Migrate the original MySQL business database to the DMP platform and perform vertical splitting to form more than 28 independent business databases.
High availability and read-write separation configuration: Configure high availability and read-write separation features for each business database to ensure zero data loss during fault switching and automatically perform read-write separation horizontal expansion when the business volume increases to improve the throughput of the database.

DTS Cluster Deployment

DTS cluster installation and configuration: Install and configure the Action Technology CloudTree DTS cluster to ensure the stable operation of the cluster.
Real-time data aggregation configuration: Configure the DTS cluster to realize the real-time aggregation of data from more than 20 business modules such as the user center, account center, order center, and payment center, and more than 40 logical databases. Ensure that the data in the production database can be synchronized to the big data platform in real time.
Big data platform docking: Docking with the big data platform of Industrial Bank to ensure that the synchronized data can be used by business personnel for data statistics and risk control analysis in real time.

Implementation and Execution

Data volume, business volume, and other situations of DTS and DMP so far:

Total data volume: 2TB
Maximum single database: 500GB
Average daily transactions: 4 million times, average daily visits: 16 million, business concurrency: 30,000
Order database: 80 million data, peak 1500TPS
Payment database: 150 million data, peak 1200TPS
Account database transaction flow data: more than 400 million, peak 36000QPS/400TPS

Achievements and Benefits

Action Technology's financial-grade high availability, distributed middleware, large-scale operation and maintenance management platform, and development and operation system specification construction services provide the bank with the ability to support 10,000+ MySQL instances to meet the needs of different business scenarios, reduce the TCO of the database, and ensure the core system. For example, during the Double 11 transaction peak, it reaches 20,000 TPS. This solution realizes the urban-level disaster recovery solution of two places and three centers, supports the RPO = 0 of the database of same-city dual-alive, meets the bank's 5-level disaster recovery requirements, and meets the 7x24-hour service requirements.

Yum China

With more than 50,000 stores worldwide and an annual turnover of over 10 billion dollars, it needed to build a highly available system database architecture. It chose Yunshu® Shard distributed database solution and Yunshu® SQLE, improving the system's high availability, stability, and operational efficiency, and ensuring the consistency and security of data.

Starbucks

With more than 38,000 stores worldwide and an annual sales volume of nearly 10 billion dollars, to enhance the high availability and data consistency of the system, it chose Yunshu® Shard distributed database solution and Yunshu® SQLE, improving the system's processing capacity and operational efficiency, and ensuring the accuracy and security of data.

McDonald's

With more than 40,000 stores worldwide and an annual operating income of over 25 billion dollars, to enhance the operational service capability and system stability, it chose the DMP multi-source database management platform and the open-source database MySQL, ensuring the stable operation of the system, reducing operational costs, and ensuring business continuity and data integrity.