Abstract: Industrial software is not won by feature checklists, but by sustained evolvability. For infrastructure software such as databases and real-time data systems—once deployed into production—long lifecycles, deep ecosystem coupling, high switching costs, and unpredictable requirements become unavoidable constraints. In this context, open source is often misinterpreted as idealism or marketing; in practice, it is a structural engineering choice: building trust through verifiable transparency, reducing vendor lock-in through open interfaces, accelerating iteration through community feedback, and defining accountability through enterprise-grade capabilities and services. This article addresses four questions practitioners care about most: why opening core code matters, what should be open, how open source becomes a business, and how open source can advance China’s industrial software. Using TDengine as a real-world case, it argues that meaningful open source is not merely “publishing code”, but exposing the core capability and evolution path—so ecosystems can plug in, users can verify, and products can be shaped by real-world use.