Google Cloud Platform (GCP)

লেসন ৩৬ / ৩৮

☁️ GCP

  • Google cloud, data/ML strong
  • Kubernetes born here
  • Clean UI + pricing
  • Free: $300 credit 90 days + always-free tier

🆚 AWS → GCP Naming

EC2          → Compute Engine
S3           → Cloud Storage
RDS          → Cloud SQL
Lambda       → Cloud Functions / Cloud Run
ECS          → Cloud Run (preferred)
EKS          → GKE
Route53      → Cloud DNS
CloudFront   → Cloud CDN
IAM          → IAM (similar)

🖥️ Compute Engine

VM similar to EC2
Sustained use discount automatic
Live migration (no reboot on maintenance)
Preemptible VMs — 80% off (like Spot)

gcloud compute instances create vm1 \
  --machine-type=e2-micro \
  --image-family=ubuntu-2204-lts

🐳 Cloud Run 🔥

GCP-এর signature service — container → HTTPS URL।

Serverless container, scale 0 → 1000
Pay per request + time
Per-request CPU allocation
Fastest container deploy anywhere

gcloud run deploy myapp \
  --image gcr.io/project/myapp \
  --region=asia-south1 \
  --allow-unauthenticated

📦 Cloud Storage

Standard / Nearline / Coldline / Archive
Lifecycle rules similar to S3
Uniform bucket-level access
Signed URL upload

🗄️ Cloud SQL

Managed MySQL, PostgreSQL, SQL Server
HA with 99.95% SLA
Read replica, automated backup
Point-in-time recovery

🔥 Firebase (GCP family)

  • Realtime Database + Firestore
  • Authentication (Google, FB, Email, Phone)
  • Hosting (static + dynamic)
  • Cloud Functions (JS)
  • FCM push notifications
  • Perfect for mobile/web app backend

☸️ GKE (Google Kubernetes Engine)

Managed K8s
Autopilot mode — fully managed (no nodes)
Standard mode — manage nodes
Best K8s experience of major clouds

📊 BigQuery 🔥

Serverless data warehouse
Petabyte scale queries in seconds
SQL interface
$5 per TB scanned
Free: 1 TB query/month, 10 GB storage

🤖 AI/ML Services

  • Vertex AI — unified ML
  • AutoML
  • Vision/Speech/Translate APIs
  • Gemini API — GCP's LLM

💰 Free Tier (always)

Compute Engine: 1 e2-micro (us regions only)
Cloud Storage: 5 GB
Cloud Functions: 2M invocations
Cloud Run: 2M req + 360k GB-sec
BigQuery: 1 TB query, 10 GB storage
Firestore: 1 GB, 50k read/day

🎓 মূল কথা ⭐

GCP = data/ML first। Cloud Run = best container deploy experience। Firebase = mobile/web backend super fast। BigQuery = analytics magic। Free tier generous।