Platform

We all use cars. Some prefer to own a car while others prefer to rent. Which model is better is an individual question.

When it comes to enterprise computing, there have long been a single operation model – to own the computing resource. With the advent of virtualization technology, data centre co-location model, the cloud became popular. Cloud computing is essentially the same delivery model as renting a car.

The table compares the two delivery models with car and computing resource as examples

Delivery Model	Own	Rent
Example	buying a car buying servers and operating on-premise	renting a car (e.g. by mileage and time) Renting computing resource (from public cloud)
Cost	one upfront cost plus on-going maintenance	usage-based billing, with potential discount on fix-term rent (leases)
How trade happens	cars are sold in dealership computing resources are sold by hardware vendors	car rental service is delivered physically at the rental agency computing resource in the cloud is delivered over high speed network, in response to API request to public cloud vendors
Pros	Cost is relatively predictable regardless of usage	subscribe to the latest technology (or model, version, security patch, etc) without additional cost
Cons	Maintenance responsibility. Technology becomes out of date	Cost can spin out of control if usage is not being monitored closely

In the cloud, it is significantly cheaper to provision computing resources. However, moving to cloud does not automatically guarantee cost effectiveness. It is important to architect cloud-based solution the right way for the best return on investment.

With cloud computing, the rental business model also means lack of user loyalty. Cloud vendors hope to lock in their users with increasing cost of transfer. Community-led initiatives such as Kubernetes aims to standardize the cloud platform and abstract the differences away from platform users. The success of Kubernetes created an entire cloud native landscape of open-source projects. In return, some of these technologies such as service mesh greatly enabled Kubernetes platform, by allowing applications to offload non-business features down to the platform. Welcome to the cloud native era!