Construction and Future Planning of the Quality Assurance Technical System at Home

In response to the fast development of Home’s business and the rise of mobile internet, artificial intelligence, and upgraded automotive consumption models, the quality assurance team seeks more effective and efficient ways to support business growth through a robust quality assurance technical system.

The technical system follows a "one horizontal, many verticals, integration" strategy: the horizontal Paas capability is provided by the Home testing cloud platform, while each business line contributes customized vertical tools, and surrounding platform capabilities are integrated to achieve standardization, process automation, and intelligence.

The system is divided into three layers: the Paas layer (Home testing cloud platform), the business layer (customized tools for each BU), and the surrounding system layer (other internal platforms).

Construction Idea

1. The Paas layer, as a horizontal capability, saves resources, avoids duplicate construction, and brings scale‑economy benefits. 2. Vertical customized tools are necessary because the Paas layer cannot fully cover all business‑specific testing needs. 3. Quality assurance activities span multiple stages, requiring integration with surrounding systems such as pipelines and other tools.

Current Status

Business Layer Various BU quality teams have built capabilities like interface automation, Mock, quality metrics, mobile automation, and other small tools, but they suffer from high customization, poor reusability, incomplete functionality, and duplicated effort.

Paas Layer The Home testing cloud already offers code quality scanning, unit testing, data construction, automation testing, and performance testing, covering both development and testing phases. Performance testing serves all business lines; code quality scanning is used by big data, information systems, and commercial platforms; automation testing is being iteratively upgraded.

Surrounding System Integration The Paas layer has integrated with the Aone pipeline for unit and smoke automation tests and is collaborating with the commercial platform quality team to embed Paas tools into business testing workflows.

Paas Capability Practice

Performance Testing Version 3.0 supports high‑QPS (100w+), stable and efficient load testing, and has been used for large events like the 818 car exhibition.

The product includes modules for creating load tests, scenarios, plans, reports, and online packet capture, supporting custom requests, traffic replay, JMeter, and gRPC.

Automation Testing Supports interface‑level and scenario‑level automation, usable for smoke, iteration, and regression testing. Currently supports HTTP/HTTPS, with plans for RPC, TCP, etc. Collaboration with the commercial team is driving smoke test adoption.

Unit Testing Evaluates pre‑release quality, supports Java services, provides coverage statistics and reports, and integrates with Aone.

Coverage and report screenshots are shown in the following images.

Code Quality Scanning Supports 20+ languages, incremental scanning, high efficiency, multiple trigger methods, and integration with Aone for continuous detection.

Integration screenshots with Aone are also provided.

Challenges

1. Delivering high‑value output from the Paas layer. 2. Enhancing productization of Paas tools. 3. Technical breadth covering front‑end, mobile, back‑end, various languages, frameworks, middleware, networking, virtualization, operations, big data, AI, etc. 4. Abstracting a universal product model for diverse business needs. 5. Resource efficiency, especially human resources.

Future Planning

Prioritize operation support and iterative optimization of existing products, fill high‑priority business needs, and gradually add missing testing capabilities. In 2022, focus on mobile testing capability while upgrading code scanning, automation, and performance testing.

Conclusion

The article presents an overview of the quality assurance technical system from a high‑level perspective, noting that detailed product information will be covered in future articles and inviting community feedback.

Author Bio Luo Ping, Cloud Platform Department, Automotive Home, responsible for testing cloud platform development.