Why Operations Won’t Die: A Veteran’s Perspective

Recently an article claimed the entire operations industry might be killed, a view that lacks basic understanding. As a veteran with over ten years of experience, I share my perspective and invite discussion.

Note 1: This piece addresses a viewpoint, not individuals. Note 2: Because a previous self‑media article spread widely, I repost to prevent misinformation and encourage operations colleagues to stay focused.

What Is Operations

Operations is a function. It emerged with the software industry and will disappear only when the software industry collapses; people may vanish, but functions persist.

Software engineering requires requirement analysis, development, testing, as well as deployment and maintenance, giving rise to operations.

Software’s lifecycle resembles human life, but development only gives birth, not nurture.

Who nurtures? Operations.

Developers often say, “It runs fine in my environment…”.

Your development environment runs alone, but production may serve thousands or millions of concurrent users.

A colleague once asked whether architects should come from development; the answer is yes—architects should stem from operations because high availability, high performance, and high scalability all start from the production perspective.

Operations Functions Beyond Deployment

Over the past 20+ years, operations has four recognized responsibilities: quality, efficiency, cost, and security (many companies’ security departments are overseen by ops directors).

Operations should thank cloud computing.

First, cloud reduces scapegoating and enables automated resource delivery.

Second, cloud forces ops to improve, allowing them to deliver higher‑value business work instead of remaining mere craftsmen.

Cloud computing is a new silo that intensifies quality and security challenges, making ops even more critical.

Without business and system disaster recovery, events like cloud‑provider outages or I/O hangs leave ops anxious.

Who coordinates? Only operations.

Efficiency is more of a business concern and less related to cloud infrastructure.

Cost‑wise, cloud is more expensive than amortized physical machines; redundancy can double expenses.

Ops can save money—Tencent’s SNG ops team saved 1 billion RMB annually through backend optimization.

Saving money is a manifestation; the core is ops’ deep mastery of technology and business, a genuine love for the company.

Means to Achieve Functions May Change

Ops has evolved from manual (hand‑crafted) ops to automated ops and now to intelligent ops.

Industry has also moved horizontally from automated ops to DevOps, following the same development law.

Human society progressed from agriculture to workshops to industrial assembly lines.

Industrial era split into mass‑production Ford lines and lean‑production Toyota lines.

Software industry was born after industrialization and is still in a painful manual‑workshop stage.

Programmers calling themselves “code farmers” is not mere self‑deprecation.

Software must follow existing laws, learn from industrialization, and build pipeline‑style production.

Agile development stems from lean industrial thinking, but it only lets development and testing collaborate to quickly produce packages (birth) without solving deployment and maintenance (raising children).

Thus DevOps emerged; a colleague sees DevOps as agile’s extension to operations.

But DevOps will not kill operations.

DevOps frees ops from routine deployment; some companies even let developers handle releases.

Good for developers—now they can jump into the pits they dug.

DevOps does not kill ops; the core remains ops’ rich functions: quality, cost, efficiency, and security.

The boundary between ops and development may blur, but it does not mean replacement.

Not Everyone Can Do Operations

Only meticulous, bold, responsible, and humble individuals can succeed in ops. When incidents occur, poor mental resilience leads to failure.

Can Cloud Outages Kill the Operations Industry?

It’s a laughable question, as a netizen remarked.

Cloud outages actually highlight the importance of ops. If a company believes external resources can solve everything, it ties its fate to others.

Harsh truth: if you don’t die, who does?

Both outage‑prone and stable clouds cannot kill the ops industry for the same reason.

Hope self‑media use their ability to avoid panic; the statement itself contains logical errors.

Operations’ Spring Is Near

You feel cold only because it seems winter.

Traditional enterprises still suffer from manual ops (e.g., script repositories).

The internet industry may only account for about 5 % of national GDP.

Ops in the internet’s red sea may feel devalued, but traditional enterprises present a vast blue‑sea market.

IoT will further increase ops demand.

A friend imagines every streetlight in Beijing equipped with IT access—massive ops opportunities.

Ops personnel need to upgrade their knowledge: learn Python for DevOps, AI for intelligent ops, and business knowledge to become technical operations.

Internal job flow has nothing to do with whether the ops industry disappears. Ops need a sense of crisis but not fear.

The sky is as big as the ops world.

Below is a vibrant discussion from the efficient‑ops community:

Ops internal job flow has nothing to do with the industry’s disappearance. The new era ops have a fresh mission—share your thoughts.