Xinjiang Rural Credit Cooperative ATM Monitoring System Construction Project

Project Overview

As banks shift to smart branches, the surge in self-service devices creates new demands on ATMC, ATMV, and ATMP systems—from cross-vendor hardware compatibility to faster business development. With multiple vendors, inconsistent interfaces, and fragmented operating systems, delivering a unified service experience is hard. Xinjiang Rural Credit Cooperatives engaged a smart banking solution provider to build cross-platform software that unifies all business functions, processes, and channel service quality—strengthening the bank's overall brand image.

Solutions

1. The database uses GBase 8s with dual-server load balancing, requiring two independent servers: the first as the primary database and the second as the backup.

2. The backend services are built using Spring Boot and Spring Cloud microservices, with Redis Sentinel for caching, RabbitMQ for message queuing, and TongWeb as the front-end container.

3. Eureka, Zuul, and the front-end service WAR packages are deployed together with TongWeb on two independent servers, each running these three services.

4. Eureka serves as the service registry and discovery server, Zuul as the API gateway; TongWeb is the web container used to deploy the front-end WAR packages.

5. When a browser accesses a page, requests go through the F5 load balancer and are then forwarded directly to the TongWeb server. For data requests to backend microservices, the browser sends them through the F5, then to the Zuul server, which finally routes them to the appropriate backend microservices.

6. The file service subsystem requires two servers, each running this service, and it must register with Eureka.

7. The V-end status and transaction message processing service requires five independent servers. Netty is used to receive terminal requests. RabbitMQ is also deployed here, with one RabbitMQ instance on each server.

8. The Redis Sentinel service is deployed on three independent servers.

Results

Implementation Details

The entire project deployed two sets of GBase 8s V8.8 to set up an HAC (High Availability) primary-standby cluster, with the primary node providing read and write services and the standby node providing read services. The project went live in 2022.

Technology Stack:

  • Processors: Hygon

  • Operating System: Kylin V10

  • Database: GBase 8s

Results & Value

  • Full Domestic Compliance: Deployed on all domestic hardware and software, meeting financial IT application innovation (Xinchuang) requirements.

  • Lower Costs: Adopting domestic database software reduces overall project costs.

  • High Availability: GBase 8s primary-standby cluster with automatic failover ensures continuous service.

  • Easy Maintenance: User-friendly interface with graphical management tools simplifies troubleshooting and reduces operational complexity.