Cloud benefits: high availability vs reliability vs scalability
Verdict: Maximum uptime backed by SLAs, high availability. Recover from failures and keep functioning, reliability. Adjust resources to match changing demand, scalability.
| Criterion | High availability | Reliability | Scalability |
|---|---|---|---|
| Concerns | Uptime through disruptions | Recovering from failures | Matching resources to demand |
| Backed by | Service-level agreements (SLAs) | Decentralised global design | Consumption-based model |
| Described as | Reachable with uptime guarantees | Recover and continue after a failure | Scale up or out with demand |
Rules
- High availability ensures maximum uptime regardless of disruptions, with the uptime levels a service promises formalised in its SLAs.
- Scalability is the ability to adjust resources to meet changing demand, so you pay only for the capacity you use.
- Reliability is the Well-Architected pillar for recovering from failures and continuing to function, aided by the cloud's decentralised design.
Traps
- Reliability concerns recovering from failures, not the uptime guarantees expressed in SLAs; those describe high availability.
- Scalability adjusts resource amounts to demand; it is not recovering from a failure, which is reliability.