What Is Cloud Native? A Simple City Analogy and Evolution Explained

Question 1: “Who am I?”

Before the holiday a friend asked me to explain cloud native in the simplest terms. I used a city analogy: the cloud is like a city, its residents are citizens, and cloud applications are the inhabitants.

Analogy between city native and cloud native

In a city, people are born, live, and act in the city; in the cloud, applications are born, live, and act in the cloud, offering faster, more agile, and more elastic capabilities.

A

City native

Cloud native

Status/Means

Born in city, grow in city

Born in cloud, grow in cloud

Target

Citizens (people)

Applications

Result/Purpose

Typical city‑resident traits

Simpler, faster, more elastic, quicker updates and deployments

Born‑native

Born in city, live as city citizens

Born in cloud, use cloud‑hosted services to achieve the result

Lift‑and‑shift

Move to city but keep old habits

Non‑cloud‑native apps lifted without refactor cannot become cloud native

Cloud‑native transformation

Move home to city and adopt city ways

Move application to cloud and refactor to cloud‑native form

The author believes the object of cloud native is the application. Narrowly, an application is a system running on the cloud; broadly, it includes all software running on cloud infrastructure, including cloud services.

Key characteristic: speed

Cloud‑native applications are fast to develop, launch, update, scale, and respond to market and user changes.

Proposed definition

A cloud‑native application runs on the cloud, fully utilizes cloud‑hosted services and technologies such as containers, Kubernetes, micro‑services, and DevOps to achieve comprehensive “fast” (development, deployment, scaling, etc.) for better business agility.

Implementation approaches

The term “cloud native” first appeared in a 2010 blog post by Paul Fremantle, following the earlier emergence of cloud computing in 2007. Early examples like Netflix leveraged AWS services and VM elasticity to achieve basic cloud‑native capabilities.

Current mainstream methods include cloud‑hosted services, containers, micro‑services, Kubernetes, and DevOps. Containers standardize application form, micro‑services split monoliths, Kubernetes orchestrates containers, and DevOps improves development and release efficiency.

Relation to Kubernetes and containers

Kubernetes is a leading technology for implementing cloud‑native applications but not sufficient alone; an application must exhibit the defined cloud‑native traits, not merely be containerized.

Cloud‑native technology stack

A simplified, non‑exhaustive diagram (see image) illustrates the stack; full stacks are documented by CNCF.

Question 2: “Where do I come from?”

The first development stage is based on cloud‑hosted services and VM elasticity. The “Wang’s Four Rules” summarize practical steps:

Rule

Practice

Value

1

Store static files in object storage

Cost‑effective, scalable, flexible storage for static data

2

Use role instead of access/secret keys

Higher security and easier management

3

Prefer managed services

Cost optimization, strong fault tolerance, good performance, rapid development

4

Do not keep data on servers

Optimized cost, fault tolerance, security, development advantages

Applications following these practices are considered to be in the first cloud‑native stage.

Question 3: “Where am I going?”

Stage 2 (around 2014) introduced containerization: Docker’s first release and Kubernetes’ first release enabled massive adoption of containers, providing an order‑of‑magnitude improvement in elasticity.

Containerized applications become standardized, easier to migrate, and can run on any platform via Kubernetes, with service meshes (e.g., Istio) moving infrastructure code to the mesh layer.

Stage 3 moves toward serverless: cloud providers offer serverless services (e.g., AWS Lambda) where users upload code and the platform handles execution, achieving true event‑driven, on‑demand operation.

Additional questions

Can cloud‑native be achieved in private clouds? The author argues that stages 1 and 2 can be realized on modern “full‑stack” private clouds that adopt public‑cloud stacks.

Do traditional enterprises need cloud native? The author contends that faster, more agile applications are essential for traditional sectors such as finance, though implementation should be staged and context‑aware.

Written during the 2023 Spring Festival return journey.