Why Kubernetes Was Born: A Story‑Driven Intro for Frontend Developers

0 Introduction

Last year I switched roles and began working with Kubernetes. My knowledge was still shallow, so I decided to share the key insights I gathered to give readers a beginner’s understanding of K8s.

1 The Story Begins

As living standards rise, car ownership expands, and new regulations on vehicle scrappage emerge, a colleague sees a business opportunity and starts a platform called “Taoche.com”.

2 Story Development

The platform initially runs as an all‑in‑one Java application on a single physical server. Rapid growth forces hardware upgrades, then a split into multiple services using middleware such as HSF, TDDL, TAIR, Diamond, and MetaQ. Virtual machines hide hardware differences but add performance overhead.

Switching to Docker reduces overhead and simplifies CI/CD, but the growing number of containers creates scheduling and networking challenges.

To address these, a container orchestration system is designed, consisting of master and worker nodes.

3 Container Orchestration System

Master nodes run management components, while worker nodes run business containers. The master provides two main interfaces: one for administrators (kubectl) and one for workers (kubelet).

The kube‑apiserver exposes the API; the kube‑scheduler decides where to place containers based on resource reports from kubelet . Persistent state is stored in etcd , which offers strong consistency and high availability.

Node components include kubelet (reports node and pod status) and kube‑proxy (handles pod‑to‑pod networking using iptables/ipvs). Services and DNS provide stable virtual IPs and names for pods, abstracting away changing pod IPs.

Additional plugins such as DNS, a web UI, resource monitoring, and log aggregation improve the user experience.

Master components: kube‑apiserver, kube‑scheduler, etcd, kube‑controller‑manager

Node components: kubelet, kube‑proxy

Plugins: DNS, web UI, container metrics, cluster logs

4 Serverless Container Orchestration

Even though the custom orchestration works, the company’s leadership seeks to reduce operational cost. They wonder whether a serverless model could let developers focus on business logic without managing clusters, and invite ideas on how to build a serverless orchestration system.