Basic Information
| Field | Details |
|---|---|
| English name | Scalability & Elasticity |
| Full name | Scalability & Elasticity |
| Chinese description | 可扩展性与弹性 |
| Japanese description | スケーラビリティと弾力性 |
| Exam frequency | ⭐⭐⭐⭐⭐ |
| Often confused with | High Availability |
In one sentence
Scalability means a system can grow; elasticity means capacity expands and contracts with demand.
Core purpose
- Scale Out / In changes instance count and suits stateless, distributed applications.
- Scale Up / Down changes one instance size, but has hardware limits and may require downtime.
- An ASG uses Min, Desired, and Max as capacity boundaries and adjusts them from metrics or schedules.
Exam focus
- Auto Scaling manages capacity while ELB manages traffic; they work together but have different jobs.
- Target Tracking follows a target metric; Scheduled Scaling fits predictable peaks.
- Scalability is not high availability; HA also requires failure-domain isolation and failover.
- Scaling plans must account for databases, service quotas, cost, and downstream throughput.
Common misconceptions
- A scalable system does not necessarily scale back automatically during quiet periods.
- Vertical scaling is not the same as automatic elasticity.
- Auto Scaling cannot repair inefficient application code or database queries.
Key takeaway
Scalability = handle growth; Elasticity = match demand; ASG = Min ≤ Desired ≤ Max.
Related services
EC2 Auto Scaling, Elastic Load Balancing, CloudWatch, SQS, ECS Service Auto Scaling.