Understanding DevOps: Integrating Development and Operations Beyond the ‘Who Develops Who Operates’ Myth

We often say DevOps means "who develops also operates" and a seamless integration of development and operations, yet few can explain how to achieve this in practice, and the notion that developers should also operate is unrealistic.

DevOps promotes development‑operations integration, not "who develops who operates". Google SRE shows that while SRE work is fundamentally operations, it involves developing operational tools that automate and intelligent‑ify routine tasks, reducing manual maintenance and shifting the capability from developers to the operations team itself.

True DevOps integration does not make developers perform operations; instead, mechanisms and processes should organically combine the two groups into a single, efficient whole, eliminating gaps and improving service development and operational efficiency.

Before or during DevOps adoption, it is essential to consider how to resolve conflicts of interest between development and operations and how to boost efficiency.

Google SRE provides valuable insights for the operations phase of DevOps. Operations staff must deeply understand their environment, tools, processes, and resources, and be able to independently develop automation tools that achieve high availability, stability, and security, supporting the full lifecycle of application operations.

SRE focuses on the Ops stage of DevOps, while the Dev stage covers requirements, design, coding, testing, and deployment, emphasizing continuous development, continuous integration, and continuous delivery to improve efficiency. Consistent and agile environments are crucial for a high‑performing Dev stage.

Traditionally, developers maintain development, test, and UAT environments, leading to inconsistencies with production and causing unexpected issues during deployment. In DevOps, environment maintenance should be centralized by the operations team, while developers simply use the environments.

Unified permission and authentication management should be provided by a technical middle‑platform operated by the ops team, offering a PaaS‑style single sign‑on that lets developers access any environment with their own accounts, avoiding duplicated effort.

Environment consistency can be partially achieved through containerization, though differences in host configurations still cause performance variations.

Preparing environments agilely is challenging; a lot of time is spent on setup. Whether using containers or not, tools and standardized processes can automate environment preparation, test data generation, test case construction, and defect recording, thereby improving efficiency.

Software coding standards are often low, relying heavily on individual developer ability; outsourcing software development for full control is generally unrealistic.

Development must ultimately deliver standardized artifacts—container images, JAR/WAR/EXE files—that are centrally managed (e.g., image registries, configuration management tools) and deployed automatically via scripts or tools to support continuous deployment.

Our "Container Cloud Platform Operations Architecture Design" explains how standardized delivery links development and operations, reducing communication costs and improving efficiency; container images act as standardized delivery media connecting development, testing, UAT, and production stages.

A closed feedback loop is vital: operations must provide developers with not only bug reports but also user‑experience insights, requiring appropriate feedback mechanisms covering both technical and non‑technical factors.

Horizontally, we can view the flow as development → standardized delivery → operations, where standardized delivery acts like a joint that lubricates and connects the two sides, allowing each to leverage its expertise. Vertically, operations can be layered: infrastructure maintenance at the base, followed by platform, tool, and environment maintenance, and finally business‑application operations, all ensuring availability and security.

From a lifecycle perspective, about 80% of work occurs in the operations phase; professionalizing, automating, and even intelligent‑ifying these tasks greatly boosts efficiency. Development‑operations integration focuses on improving operational efficiency across applications, environments, platforms, tools, and infrastructure, ensuring high availability and stability akin to Google SRE’s site reliability.

DevOps integration seeks aligned interests between development and operations, not a single person doing both; mechanisms and appropriate tools must provide the “joint” and “lubricant” that bind the two functions together.