Operations Engineering Evolution

It is a well known fact scaling is hard, be it in people, process or technology. In hypergrowth markets, it is well understood that processes will break every 6 months because the scale is changing fast. Same is the case in infrastructure and technical operation space. Lets cover how the area of Reliability Engineering has evolved over time.

Evolution of Operations

It all started with "How do we streamline Technical Operations", it has gone through a lot of iterations

Ops -> Application Operations (appops, sysops) -> DevOps -> SRE ->NoOps -> GitOps

Each of them attacked a problem area and solved for that. Lets go over the journey (starting 2000):

Ops (2000-2004) : Responsible for server procurement, provisioning and OS/Network maintenance

AppOps (2004-2012): Got closer to the Application stack, started to own JVM, OS fine tuning for the app, logging/monitoring and tier-1 support.

DevOps (2012-current): Brought Dev and Ops together. Gave rise to the concept of CICD. Ops started to own the automated pipelines and moved all ops activities to git/SCM. Later in the game Infra as Code/ Infra is code became popular and that became the standard practice.

SRE* : CICD became almost a solved problem, everyone had a flavor of CICD but did not solve reliability of the systems, the focus shifted towards advanced deployments ( blue-green, canary, feature flagging) to help with ensuring the reliability (uptime, availability, resiliency, proactive detection and remediation)

SRE* + PE**: Kudos to Google and Facebook for leading the journey and establishing a lot of best practices around how to scale effectively and efficiently. They also led the front around how orgs should be structured to support operations as massive scale. This solve for ensuring right tools to address scale, reliability and best practices around massive operations at scale.

• SRE* = Site Reliability Engineering
• PE** = Production Engineering

Feel free to share where you are on your journey and learnings from it.