Horizontal scaling, or scaling out, increases capacity by adding machines or application instances that share the load. Compared with vertical scaling, it offers greater elasticity and fault tolerance, but requires more orchestration around load balancing and state management. section 6.3
Use the following comparison when reasoning about architecture or answering interview follow-ups:
- Choose vertical scaling for simpler monolithic systems, limited growth, or situations where refactoring is not feasible. section 6.3
- Choose horizontal scaling for distributed, cloud-native systems where resilience, elasticity, and continued capacity are priorities. section 6.3
- Pair multiple instances with load balancing so traffic can be distributed across available machines. section 6.3
- Address distributed state management explicitly because adding instances makes coordination more complex. section 6.3
Table 6.4 Scaling strategies
|
Topic
|
Fundamentals
|
Applied skills
|
Advanced
|
|
Vertical vs. horizontal scaling
|
Understands that vertical scaling adds CPU/memory to a single node, while horizontal adds nodes.
|
Chooses horizontal scaling for stateless, cloud-native apps.
|
Designs for distributed state management and consistent performance under dynamic scaling.
|
|
Load balancing
|
Knows the role of load balancers (LB) in distributing traffic.
|
Implements NGINX, HAProxy, or cloud ELBs to balance traffic across instances.
|
Tunes load balancers for sticky sessions, weighted routing, failover, and cross-region balancing.
|
|
Autoscaling & elasticity
|
Understands that autoscaling adjusts resources based on load.
|
Configures autoscaling policies using metrics like CPU, memory, or queue depth.
|
Designs for elasticity across microservices with predictive scaling, cost optimization, and graceful scale-in/out mechanisms.
|
Compare vertical and horizontal scaling across fundamentals, applied choices, and advanced concerns such as distributed state, load-balancer tuning, and elasticity. section 6.3
Interview scenario
When asked to explain the difference between scaling strategies, define vertical scaling as enlarging one machine and horizontal scaling as adding machines. Then justify horizontal scaling for a stateless cloud-native application and explain how a load balancer distributes requests across instances. section 6.3


















