Google Cloud Platform (GCP) is Google's public cloud computing platform offering on-demand compute, storage, networking, databases, machine learning, and managed services that run on the same infrastructure powering Google Search, YouTube, and Gmail. GCP competes with AWS and Azure by emphasizing deep AI/ML integration (Gemini models, Vertex AI Agent Builder), global networking performance, and serverless-first architectures. A key mental model: GCP services are organized around a resource hierarchy (organizations → folders → projects → resources) with IAM for fine-grained access control, making multi-project, multi-team deployments easier to govern at scale than flat permission models.
What This Cheat Sheet Covers
This topic spans 20 focused tables and 114 indexed concepts, 116 flashcards. Below is a complete table-by-table outline of this topic, spanning foundational concepts through advanced details.
A jump-to index of every table row in this cheat sheet.
An interactive map of every table and concept in this topic.
Table 1: Compute Services
Compute is where your code actually runs, and GCP gives you a full spectrum of options — from raw virtual machines you fully control, through managed Kubernetes and serverless containers, down to event-driven functions that scale to zero. The right choice usually comes down to how much of the underlying infrastructure you want to manage versus hand off to Google, with Cloud Run and Cloud Functions sitting at the most hands-off end and Compute Engine at the most flexible.
| Service | Example | Description | |
|---|---|---|---|
gcloud compute instances create my-vm --zone=us-central1-a --machine-type=e2-medium | • Virtual machines (VMs) on Google infrastructure • choose predefined or custom machine types across families (E2, N2, C3, C4, H4D), supports GPUs and persistent disks | ||
gcloud run deploy my-service --image gcr.io/project/image --platform managed | • Serverless container platform that scales HTTP containers from zero to thousands • pay-per-request, supports GPU workloads, websockets, and gRPC | ||
gcloud container clusters create my-cluster --num-nodes=3 | • Managed Kubernetes for container orchestration • Autopilot mode (Google manages nodes) and Standard mode (you manage nodes) | ||
gcloud app deploy app.yaml | • Platform-as-a-Service (PaaS) for web apps and APIs • Standard environment (sandboxed, fast startup) or Flexible environment (custom runtimes on VMs) | ||
gcloud functions deploy myFunc --runtime python312 --trigger-http --gen2 | • Serverless event-driven functions triggered by HTTP, Pub/Sub, or Cloud Storage events • 2nd gen runs on Cloud Run infrastructure with longer timeouts | ||
gcloud batch jobs submit my-job --location=us-central1 --config=job.json | • Fully managed batch processing for HPC, ML training, and data-processing workloads • schedules and queues jobs on Compute Engine VMs with auto-provisioning | ||
gcloud workstations create my-ws --config=my-config --cluster=my-cluster | • Managed cloud-based development environments with preconfigured IDEs • runs on Compute Engine with persistent disks, supports custom container images | ||
gcloud compute instances create my-vm --provisioning-model=SPOT | • Deeply discounted VMs (60–91% off) that can be preempted when capacity is needed • ideal for fault-tolerant batch jobs, rendering, and CI/CD workloads |