Exercise: The Platform Map and Decision Guide

Build the mental model

This closing lesson does not teach a new platform; it organizes everything the course covered into one map you can return to later.

Category	Example Products & Capability Range
Major Cloud Providers	AWS, Google Cloud, Azure - full infrastructure range, from raw VMs to managed everything.
Web/Edge Platforms	Cloudflare, Vercel, Netlify - web delivery, serverless, and edge function range.
Application Platforms	Railway, Render - backend services, databases, and background job range.
Backend-as-a-Service	Supabase, Firebase - database, auth, storage, and realtime range behind a client SDK.

A map, not a strict taxonomy

The categories genuinely overlap - a web/edge platform can host a database, and a major cloud can offer BaaS-like services. Vendors keep adding features across category lines because customers want fewer platforms to manage, not more.

What matters is the reasoning habit, not memorizing which vendor offers which feature - ask what you need first, find the category, then pick a specific vendor second.

text

THE PLATFORM MAP
----------------
MAJOR CLOUDS                   WEB / EDGE PLATFORMS
(AWS, Google Cloud, Azure)     (Cloudflare, Vercel, Netlify)
range: full infra control,     range: web delivery, edge
VMs, networking, IAM           functions, serverless, CDN

APPLICATION PLATFORMS          BACKEND-AS-A-SERVICE (BaaS)
(Railway, Render)              (Supabase, Firebase)
range: backend services,       range: managed DB, auth,
containers, managed DBs        storage, realtime, low ops

Categories overlap in real products - this is a starting
map, not a strict taxonomy.

Connect it to a real scenario

The platformCategoryForNeed function below encodes the map as a lookup, the same pattern used in the first exercise. Feed it a plain description of what you need, and it returns the matching category with representative products.

need full infrastructure control -> Major Cloud Provider
need to deploy a modern frontend fast -> Web/Edge Platform
need a backend service with a database -> Application Platform
need auth, database, and realtime with minimal setup -> Backend-as-a-Service

Worked answer key (reason first, then read)

Static docs site + contact form -> Web/Edge Platform (a small serverless function covers the form). Postgres admin dashboard, no backend staff -> BaaS or Application Platform (depends on how much schema control is wanted). Custom networking + audit logs + IAM -> Major Cloud Provider (those controls exist only at that layer).

Four categories - major clouds, web/edge platforms, application platforms, and BaaS - cover most of what you'll encounter, but they overlap in real products.
Start from what you actually need (infrastructure control, frontend delivery, backend service, or ready-made backend) and let that point to a category before you pick a vendor.
Revisit the Platform Decision Guide and glossary whenever you scope a new project instead of re-deriving the reasoning from scratch.

Try the working example

javascript

function platformCategoryForNeed(need) {
  const map = {
    "need full infrastructure control": "Major Cloud Provider (e.g. AWS, Google Cloud, Azure)",
    "need to deploy a modern frontend fast": "Web/Edge Platform (e.g. Cloudflare, Vercel, Netlify)",
    "need a backend service with a database": "Application Platform (e.g. Railway, Render)",
    "need auth, database, and realtime with minimal setup": "Backend-as-a-Service (e.g. Supabase, Firebase)"
  };

  return map[need] || "Not in the starter map - re-check the course's Platform Map lesson.";
}

const needs = [
  "need full infrastructure control",
  "need to deploy a modern frontend fast",
  "need a backend service with a database",
  "need auth, database, and realtime with minimal setup"
];

for (const need of needs) {
  console.log(need, "->", platformCategoryForNeed(need));
}

You should see

The loop prints one line per need: 'need full infrastructure control -> Major Cloud Provider (e.g. AWS, Google Cloud, Azure)', 'need to deploy a modern frontend fast -> Web/Edge Platform (e.g. Cloudflare, Vercel, Netlify)', 'need a backend service with a database -> Application Platform (e.g. Railway, Render)', and 'need auth, database, and realtime with minimal setup -> Backend-as-a-Service (e.g. Supabase, Firebase)'.

5-minute try-it

Three final scenarios pull together everything from the course. Place each one into a Platform Map category and write down the specific reason for your choice.
1. A three-person team wants to ship a simple static documentation site for their open-source library, with a simple contact form that emails them when someone submits it.
2. An internal-tools team at a mid-size company needs an admin dashboard backed by a Postgres database, built by developers with no dedicated backend or DevOps background.
3. A financial-services company needs to run a customer-facing API with strict requirements: custom private networking, detailed audit logs, and fine-grained IAM policies for every service that touches customer data.
For each scenario, name the category, then name at least one specific platform from the course that fits, and explain what about the scenario ruled out the other three categories.

One important caution

Treating the Platform Map as a rigid decision tree instead of a starting orientation, and ignoring real overlap between categories.

Picking a category based on which vendor is trendiest rather than matching the category's actual capability range to the project's real requirements.

Google Cloud: What is multicloud? — Cloud Providers & Platforms

Cloud Providers & Platforms Glossary — Common Terms

Term	Meaning
Cloud Computing	Renting computing resources (servers, storage, databases) over the internet from a provider instead of owning physical hardware.
Compute	The raw processing power (CPU/GPU/memory) a cloud provider rents out to run your code.
IaaS	Infrastructure-as-a-Service: you rent raw virtual machines, storage, and networking, and manage the operating system and everything above it yourself.
PaaS	Platform-as-a-Service: the provider manages the operating system and runtime, and you just deploy your application code.
SaaS	Software-as-a-Service: you use a complete, ready-made application over the internet without deploying or managing anything yourself.
Region	A specific geographic location where a cloud provider runs its data centers, chosen to reduce latency or meet data-residency rules.
Availability Zone	An isolated data center (or group of them) within a region, used so a failure in one zone doesn't take down your whole application.
Cloud Cost Structure	The pay-as-you-go pricing model where you're billed for exactly what you use (compute time, storage, bandwidth) instead of a flat fee.
Free Tier	A limited amount of usage a provider offers at no cost, useful for learning and small projects but easy to exceed unexpectedly.
Egress	Data leaving a cloud provider's network, which is often the cost item that surprises people because inbound data is usually free but outbound is not.
AWS	Amazon Web Services, the largest major cloud provider, offering the broadest range of infrastructure and managed services.
Google Cloud	Google's major cloud provider, notable for data/analytics services and Kubernetes, which it originally created.
Azure	Microsoft's major cloud provider, often chosen by organizations already invested in Microsoft's enterprise software.
Cloud-Native	Designing an application to take advantage of cloud features (like scaling and managed services) from the start, rather than treating the cloud as just another server.
Cloudflare	A web/edge platform known for its global CDN, DNS, and edge-function products (Workers).
Edge Network	A globally distributed set of servers that run code or cache content close to users, reducing latency compared to one central data center.
Vercel	A web/edge platform built around fast frontend deployment (especially Next.js), with instant preview deployments per pull request.
Netlify	A web/edge platform similar to Vercel, focused on static sites, frontend frameworks, and serverless functions.
Preview Deployment	A temporary, shareable live version of an app built automatically from a branch or pull request, used to review changes before merging.
Railway	An application platform that deploys backend services, databases, and background workers with minimal configuration.
Render	An application platform similar to Railway, offering managed web services, databases, and background jobs.
Modern Deployment Platform	An umbrella term for platforms (like Railway, Render, Vercel, Netlify) that automate build, deploy, and scaling steps that used to require manual server setup.
BaaS	Backend-as-a-Service: a hosted backend (database, authentication, storage, realtime) you call directly from client code without writing your own server.
Supabase	An open-source BaaS built on Postgres, providing a database, auth, storage, and realtime subscriptions out of the box.
Firebase	Google's BaaS, offering a NoSQL database (Firestore), authentication, storage, and hosting, popular for mobile apps.
Row-Level Security	Database rules that restrict which rows a given user can read or write, enforced by the database itself rather than application code.
Rolling Deployment	A deployment strategy that replaces instances of the old version with the new one gradually, keeping the app available throughout.
Blue-Green Deployment	A deployment strategy that runs two identical environments (old and new) and switches all traffic over at once, enabling instant rollback.
Canary Deployment	A deployment strategy that routes a small percentage of traffic to the new version first, to catch problems before a full rollout.
Managed Container Service	A cloud service that runs your containers (like Docker images) while handling the underlying servers, scaling, and orchestration for you.
Kubernetes	An open-source system for orchestrating containers at scale across many machines, commonly available as a managed service on major clouds (brief refresher).
Vendor Lock-In	Becoming so dependent on one provider's specific features that switching to another provider becomes expensive or difficult.
Multi-Cloud	Deliberately using more than one cloud provider, often to reduce vendor lock-in or to use each provider's specific strengths.
Observability	The ability to understand what's happening inside a running system from its external outputs - logs, metrics, and traces.
Vibe Coding	Building an app largely by describing what you want to an AI tool rather than writing most of the code by hand.
Production Readiness	The checklist of concerns (security, monitoring, backups, error handling) an app should satisfy before real users depend on it.

Platform Decision Guide

If you need	Choose
Need a simple marketing/portfolio site	Start with a static hosting/CDN category platform (Cloudflare Pages, Netlify, Vercel's static hosting) - fast, usually free at this scale, and needs no server management.
Need to deploy a modern frontend app (React, Next.js, etc.)	A managed web/edge platform (Vercel, Netlify, Cloudflare) is the usual starting point - it handles builds, previews, and CDN distribution for you.
Need a full-stack framework app (frontend + API routes together)	A web/edge platform with serverless functions (Vercel, Netlify) or an application platform (Railway, Render) both fit - the choice often comes down to how much backend logic you're running.
Need a Node.js or Python backend service	An application platform (Railway, Render) is usually the fastest starting point - deploy the service directly without configuring servers yourself.
Need a long-running backend process (queue worker, websocket server)	Favor a container/VM-based option or an application platform's dedicated worker/service type - pure serverless functions are built to start and stop, not run continuously.
Need a managed database	A managed database category product (from a major cloud, an application platform's add-on, or a BaaS) saves you from running your own database server.
Need auth + database quickly with minimal setup	A BaaS (Supabase, Firebase) may fit well here - it bundles database, authentication, and often storage behind one client SDK.
Need execution close to users around the world (low latency)	An edge platform's edge-function category (Cloudflare Workers and similar) is designed exactly for this, running code near the user instead of one central region.
Need full infrastructure control (custom networking, compliance, IAM)	A major cloud provider's raw VMs/containers is the category built for this - you trade convenience for direct control over every layer.
Need to avoid heavy vendor lock-in	Favor standard technologies across categories - plain SQL over a proprietary query API, containers over a platform-specific build system, standard protocols over vendor-only SDKs where practical.
Deploying an app that was mostly AI-generated (vibe coded)	Run it through a production-readiness checklist before real users touch it, regardless of which platform category you deploy to - the category doesn't replace that review.