The Google Cloud Infrastucture has three layers:

1. Networking and Security Layer (Base layer);
2. Compute and Storage. They are decoupled so they can scale independently based on need;
3. Data and AI LM products.

Compute

Google offers a range of computing services to accomodate different needs and scales.

• Compute Engine: it is an IaaS (Infrastucture as a services) that provides compute, storage and network resources similar to a physical machine.
• Google Kubernete Engine (GKE): it runs contenerized applications in a cloud environment.
• App Engine (PaaS): Platform as a Service bind codes to libraries and is focused on application logic. It allows you to easily run and serve a web app

Cloud Run

A fully managed compute platform that enables you to run requests or event-driven stateless workloads without having to worry about servers.

It abstract platform managements, and can automatically scale up and down.

It charges you only for the resources you use.

Cloud Run Functions

Functions that are executed in response to events. It completely serverless execution environment.

Storage

Storage is decoupled from Compute and can scale separately

GCP offers fully managed storage solutions like:

• Cloud Storage: best for Unstructured Data, like standard storage of hot data (frequently accessed), nearline storage (once per month), coldline storage (once every 90 days) and archive storage (once a year)
• BigTable: Best for analytical structured data Structured Data without SQL
• CloudSQL: Best for transactional structured data Structured Data with SQL at local scalability
• Spanner: Best for transactional structured data Structured Data with SQL at global scalability
• FireStorm: Best for transactional structured data Structured Data without SQL
• BigQuery: Best for analytical structured data Structured Data with SQL