AWS Knowledge Hub

English

Tiếng Việt

English

Tiếng Việt

Appearance

EC2 Exam Notes ​

Summary ​

  • EC2 features heavily in AWS exams (SAA, DVA, SOA, SAP); key topics include instance types & pricing models, storage, networking/security, high availability, and cost optimization.
  • Questions are scenario‑based and expect you to apply Well‑Architected principles (secure, reliable, performant, cost‑efficient) rather than memorize details only.
  • Knowing the “AWS‑preferred” defaults (IAM Role, Security Group, Multi‑AZ, Auto Scaling, EBS encrypted, Spot for batch, etc.) helps quickly eliminate wrong answers.

Exam topics mindmap ​

mermaid
mindmap
  root((EC2 Exam Topics))
    Instance Types
      Families & use cases
      Naming convention
      Burstable (T3/T4g)
    Pricing
      On-Demand
      Reserved / Savings Plans
      Spot
    Storage
      EBS volume types
      Instance store
      Snapshots
    Networking & Security
      Security Groups
      ENI & Elastic IP
      Placement Groups
    HA & Scaling
      Multi-AZ
      Auto Scaling
      Health checks
    Cost Optimization
      Right-size
      Stop/terminate unused
      Cost tools

Exam Notes by topic ​

  • Instance Types & Naming

    • Families: General Purpose (M, T), Compute Optimized (C), Memory Optimized (R, X, High Memory), Storage Optimized (I, D, H), Accelerated (P, G, Inf, Trn).
    • Naming: [family][generation].[size] (for example, m6i.large, c7g.xlarge).
    • T2/T3/T4g are burstable, use CPU credits; current‑generation types (M6/M7, C6/C7, R6/R7) are preferred over M4/C4/R4.

  • Pricing Models

    • On‑Demand: pay per hour/second, no commitment, highest cost; use for variable, short‑term, or test workloads.
    • Reserved Instances / Savings Plans: 1–3‑year commitments, save up to ~72%; best for predictable, steady‑state workloads.
    • Spot Instances: up to ~90% discount, can be interrupted with 2‑minute notice; use for fault‑tolerant/batch/CI/ML training.

  • Storage

    • EBS volume types:
      • gp3 (general purpose SSD, baseline 3,000 IOPS, 125 MB/s; configurable up to 16,000 IOPS and 1,000 MB/s).
      • io1/io2 (Provisioned IOPS SSD; io2 supports higher durability and IOPS, multi‑attach).
      • st1 (throughput‑optimized HDD for big data, logs, sequential I/O).
      • sc1 (cold HDD for infrequently accessed data).
    • Instance store: local, ephemeral, very fast; data is lost on stop/terminate; use for cache/scratch/temporary data, not for critical data.
    • Snapshots: incremental backups of EBS volumes, stored in S3 (hidden), can be copied across Regions/accounts and used to create new volumes.

  • Networking & Security

    • Security Groups: stateful, instance‑level firewalls; deny‑all inbound / allow‑all outbound by default; allow rules based on IP/CIDR or other SGs.
    • ENI: Elastic Network Interface – multiple ENIs per instance, each with its own private IPs, Security Groups, and optional Elastic IP.
    • Enhanced networking: provides higher bandwidth, lower latency, more PPS; based on SR‑IOV (ENA) or EFA for HPC/ML.

  • Placement Groups

    • Cluster: low latency, high throughput, single AZ; for tightly coupled HPC/ML workloads.
    • Spread: high availability, instances on distinct hardware; up to 7 instances per AZ; for small sets of critical instances.
    • Partition: for large distributed systems (Hadoop, Cassandra, Kafka); up to 7 partitions per AZ; isolates failures by partition.

  • High Availability & Scaling

    • Multi‑AZ deployment: spread instances across AZs, use Auto Scaling and load balancers; design for AZ failures.
    • Auto Scaling components: Launch Template/Configuration, Auto Scaling Group, scaling policies (Target Tracking, Step, Simple), and health checks.
    • Status checks:
      • System status → host issues, fix by stop/start.
      • Instance status → OS/app/network configuration issues.

  • Security

    • IAM Roles for EC2: AWS‑recommended method to grant AWS API access; temporary credentials, no static keys on instances.
    • Encryption: EBS encryption at rest with KMS, encrypted snapshots; TLS/HTTPS for data in transit, VPN/Direct Connect for secure connectivity.

  • Cost Optimization

    • Key strategies: right‑size instances, use RIs/Savings Plans for predictable usage, Spot for fault‑tolerant workloads, Auto Scaling for elasticity, and stop/terminate unused resources.
    • Cost tools: Cost Explorer, AWS Budgets, Cost Allocation Tags, RI/Savings Plans recommendations, Cost Anomaly Detection.

Exam Tips ​

  • Read each scenario carefully, highlight clues like steady usage, spiky traffic, must minimize cost, must be highly available, cannot tolerate downtime, must be fault‑tolerant.
  • Default to AWS best practices: IAM Roles, Security Groups, Multi‑AZ + Auto Scaling + load balancer, EBS encryption, CloudWatch monitoring, DR with snapshots/AMIs.
  • Eliminate options that:
    • Expose ports widely without justification (e.g. 0.0.0.0/0 to SSH/RDP).
    • Hard‑code or store access keys on instances.
    • Use single instances in a single AZ for production high‑availability scenarios.
  • Always balance cost, performance, and availability; the correct answer rarely maximizes just one of these.