Vertical and Horizontal Scaling

Choosing the Right Scaling Approach

Explanation
Scalability is the ability of a system to handle increased load without degradation in performance. The two primary scaling approaches are vertical scaling and horizontal scaling. Vertical scaling increases the resources of a single machine, such as CPU or memory, while horizontal scaling adds more machines to distribute the load.

From a system design perspective, horizontal scaling is generally preferred in modern architectures because it improves fault tolerance and supports elastic growth. However, it introduces challenges such as data partitioning, coordination, and consistency. In 2026, most scalable systems are designed to be horizontally scalable from the beginning, using stateless services and distributed data stores to simplify growth and recovery.

Table

Example
A startup initially runs its application on a single powerful server to minimize operational overhead. As user traffic grows, the system is redesigned to use multiple stateless servers behind a load balancer. This shift allows the application to scale dynamically during traffic spikes and remain available even if one server fails, demonstrating the long-term benefits of horizontal scaling.

Use Cases

Supporting rapid user growth

Handling unpredictable traffic patterns

Designing fault-tolerant systems

Building cloud-native applications 

Load Balancing Techniques

Algorithms and Load Distribution

Explanation
Load balancing distributes incoming traffic across multiple servers to prevent overload and improve availability. It acts as a control point that ensures no single server becomes a bottleneck. System designers must choose load balancing strategies based on traffic patterns, server capabilities, and application requirements.

Common algorithms include round-robin, least connections, and hash-based routing. In 2026, intelligent load balancers also incorporate health checks, auto-scaling signals, and geographic routing. Proper load balancing is critical for maintaining low latency and high reliability in distributed systems.

Table

Example
An online ticketing platform uses a load balancer with health checks to route traffic only to healthy servers. During peak booking times, new server instances are added automatically, and the load balancer begins routing traffic to them immediately. If a server fails, traffic is rerouted without user impact, maintaining system availability under heavy load.

Use Cases

High-availability web services

Traffic management during peak events

Supporting auto-scaling infrastructure

Reducing system bottlenecks