⚡ AWS CLF-C02 — Cram Jam

Last-minute dense recall sheet · read fast, stash in memory. Not a substitute for the full guide.
65 Q · 90 min · pass 700/1000
50 scored + 15 unscored · $100 · 3-yr valid
Concepts 24% · Security 30% · Tech 34% · Billing 12%
Full guide

Shared Responsibility Model

  • AWS = security OF the cloud (hardware, facilities, network, hypervisor, managed-service internals).
  • Customer = security IN the cloud (data, IAM, OS/patches on EC2, firewall config, encryption choices).
  • EC2 = customer patches guest OS/app. S3 = AWS handles infra, customer sets bucket policy/encryption/access. RDS = AWS patches engine/host, customer manages access + data. Lambda = AWS runs everything but the code and its permissions.
  • More managed → more AWS responsibility (Lambda > RDS > EC2).

Well-Architected 6 Pillars

  • Operational Excellence · Security · Reliability · Performance Efficiency · Cost Optimization · Sustainability.
  • Op Excellence = run/monitor, gain insight, improve process. Sustainability = minimize environmental impact.

6 Advantages of Cloud

  • Trade capex for variable expense.
  • Benefit from massive economies of scale.
  • Stop guessing capacity.
  • Increase speed and agility.
  • Stop spending on data centers, focus on business.
  • Go global in minutes.

Deployment Models & 7 Rs

  • Models: all-in cloud, hybrid (mix on-prem + cloud), on-premises ("private cloud").
  • 7 Rs of migration: Retire, Retain, Rehost (lift-and-shift), Relocate (VMware Cloud on AWS), Repurchase (drop & shop, e.g. SaaS), Replatform (lift-tinker-shift), Refactor/Re-architect (rebuild cloud-native).

Global Infrastructure

  • Region = isolated geographic area, multiple AZs.
  • Availability Zone (AZ) = 1+ discrete data centers, own power/net, ≥3 per Region, low-latency links between AZs.
  • Edge location = CloudFront/Route 53 caching, more sites than Regions.
  • Local Zone = compute/storage closer to a metro for low latency. Wavelength = compute at 5G telco edge. Outposts = AWS hardware in your own data center.

IAM

  • Global service, free. Default = implicit deny; explicit Deny beats any Allow.
  • Users = people/apps. Groups = collections of users. Roles = temp permissions, assumed, no long-term creds. Policies = JSON permission documents attached to any of the above.
  • Root account: lock away, enable MFA, don't use day to day, don't create access keys for it.
  • Best practice: least privilege, use roles for EC2 not hardcoded keys, enable MFA everywhere.

Organizations & SCPs

  • Organizations = manage many accounts centrally, consolidated billing (one bill, combined volume discounts), OUs to group accounts.
  • SCP (Service Control Policy) = account-level guardrail, sets max permissions, does NOT grant permissions by itself.
  • Control Tower = automated multi-account landing zone with guardrails.

Security Services (by job)

JobService
DDoS protectionShield (Std free, Adv $)
Web app filtering (SQLi/XSS)WAF
Threat detection from logsGuardDuty
Vulnerability/CVE scanningInspector
Find PII/sensitive data in S3Macie
Investigate root causeDetective
Aggregate security findingsSecurity Hub
Track resource config/complianceConfig
Log every API call (who/what/when)CloudTrail
Download compliance reportsArtifact

Encryption & Key Management

  • KMS = managed encryption keys, multi-tenant, integrates with most AWS services.
  • CloudHSM = dedicated, single-tenant hardware security module, you manage keys, FIPS 140-2.
  • Encryption at rest (S3/EBS/RDS) and in transit (TLS/SSL) both matter; ACM issues free TLS certs.

EC2 Purchasing Options

OptionBest for
On-Demandshort/unpredictable, no commitment (priciest per hour)
Reserved (1-3yr)steady-state, predictable, up to ~72% off
Savings Plansflexible $/hr commitment across instance families
Spotfault-tolerant/flexible, up to ~90% off, can be reclaimed
Dedicated Host/Instancecompliance, BYOL, physical isolation (priciest)

Cost order: Spot < Reserved/Savings Plans < On-Demand < Dedicated.

S3 Storage Classes

ClassRetrieval
Standardfrequent, ms, multi-AZ
Intelligent-Tieringunknown/changing pattern, auto-moves
Standard-IAinfrequent, ms, multi-AZ
One Zone-IAinfrequent, ms, single AZ (cheaper, less durable)
Glacier Instant Retrievalarchive, ms
Glacier Flexible Retrievalarchive, minutes-hours
Glacier Deep Archivecheapest, ~12 hours

Storage Picker

  • S3 = object storage, internet-accessible, durable, unlimited scale.
  • EBS = block storage, attaches to one EC2, single-AZ.
  • EFS = managed NFS file storage, shared across many EC2, multi-AZ.
  • FSx = managed Windows File Server or Lustre (HPC) file systems.
  • Instance Store = ephemeral, physically attached, fastest, lost on stop.

Database Picker

  • RDS = managed relational (MySQL/PostgreSQL/MariaDB/Oracle/SQL Server).
  • Aurora = AWS relational, MySQL/PostgreSQL compatible, faster + more resilient.
  • DynamoDB = serverless key-value/NoSQL, single-digit ms at scale.
  • Redshift = data warehouse, OLAP, analytics/BI at scale.
  • ElastiCache = in-memory cache (Redis/Memcached), speed up reads.
  • Neptune = graph database (relationships). DocumentDB = MongoDB-compatible document store.

Compute Picker

  • EC2 = virtual machines, full control, you manage OS.
  • Lambda = serverless functions, event-driven, pay per invocation/duration.
  • ECS/EKS = container orchestration (AWS-native / Kubernetes).
  • Fargate = serverless compute for containers, no servers to manage.
  • Elastic Beanstalk = upload code, AWS handles provisioning/scaling/monitoring (PaaS).
  • Lightsail = simplified VPS, flat pricing, small/simple apps.

Networking

  • VPC = your isolated virtual network, subnets carve it up per AZ.
  • Security Group = instance-level, stateful, allow rules only. NACL = subnet-level, stateless, allow + deny, numbered order. (Classic trap: stateful vs stateless.)
  • Route 53 = managed DNS + domain registration + health checks/failover routing.
  • CloudFront = CDN, caches content at edge locations, cuts latency + origin load.
  • Direct Connect = dedicated private physical link to AWS, consistent bandwidth, setup takes weeks. Site-to-Site VPN = encrypted over internet, fast setup, less consistent.
  • Global Accelerator = static anycast IPs, routes traffic to nearest healthy endpoint over AWS backbone.

App Integration

NeedUse
Decouple, buffer/queueSQS
Fan-out notify many subscribersSNS
Route events between services/SaaSEventBridge
Coordinate multi-step workflowStep Functions

Monitoring & Management

  • CloudWatch = metrics, logs, alarms, dashboards (performance/health).
  • CloudTrail = API call history/audit log (who did what, when).
  • Config = tracks resource configuration changes + compliance over time.
  • Trusted Advisor = automated best-practice checks: cost, performance, security, fault tolerance, service limits.
  • Systems Manager = patch, run commands, parameter store, operational visibility at scale.
  • CloudFormation = infrastructure as code, provisions resources from templates.

Billing & Cost Tools

  • Cost Explorer = visualize + analyze historical/forecasted spend.
  • Budgets = set thresholds, alert when spend/usage exceeds them.
  • Cost and Usage Report (CUR) = most granular, detailed billing data to S3.
  • Pricing Calculator = estimate costs before deploying.
  • Billing Conductor = customize billing/pro-forma pricing for customers (resellers).
  • Consolidated billing (Organizations) = one bill, shared volume discounts/Savings Plans.
  • Free Tier flavors: Always free, 12-months free, Trials (short-term).

Support Plans

PlanResponse (critical)ContactTAM
Basicn/a (self/docs/forums only)noneno
Developer<24h (general guidance)business hours emailno
Business<1 hr (prod system down)24/7 phone/chat/emailno
Enterprise On-Ramp<30 min24/7, pooled TAMpooled
Enterprise<15 min24/7, conciergededicated

Full Trusted Advisor checks (all categories) start at Business; Enterprise adds a dedicated TAM. Note: AWS is retiring Enterprise On-Ramp on 1 Jan 2027, but it is still on the CLF-C02 exam guide.

Classic Traps

SCP ≠ grant permission, only limits · SG stateful vs NACL stateless · Direct Connect (private, slow setup) vs VPN (encrypted, fast setup) · CloudTrail (who did what) vs CloudWatch (metrics/health) vs Config (config history) · KMS (managed, multi-tenant) vs CloudHSM (dedicated, FIPS) · Basic support has NO technical contact at all.
CLF-C02 rewards the simplest, most managed, cheapest answer when several look technically correct. It also tests knowing who (AWS vs customer) is responsible, not deep how-to.
Cram Jam · AWS CLF-C02 · pairs with the full study guide · verify limits/pricing against current AWS docs · not affiliated with AWS.