Cluster

A cluster is a group of two or more servers connected over a network and configured so that, from the user's perspective, they act as a single coherent system. The individual machines that make up a cluster are called nodes. The goal of this arrangement is to eliminate any single point of failure and to handle more traffic than one server could bear alone.

How a cluster works

The nodes of a cluster continuously communicate with one another, exchanging status information (a so-called heartbeat). This lets the cluster know which machines are healthy and react to failures. Depending on the goal, there are several basic types:

• high-availability (HA) cluster — if the active node fails, a standby immediately takes over its tasks (failover), minimizing downtime;
• load-balancing cluster — traffic is distributed across the nodes by a load balancer, increasing throughput;
• compute (HPC) cluster — nodes combine their power to jointly perform complex calculations;
• database cluster — data is replicated across nodes, providing fault tolerance and read scaling.

A key mechanism here is replication — synchronizing data and state between nodes so that a takeover of tasks happens without data loss.

Practical application

Clusters are the foundation of services that must run without interruption: online stores at sales peaks, SaaS platforms, banking systems and large databases. Thanks to them, updates and maintenance can proceed without taking the service offline — nodes are drained and updated one at a time. This is also the basis of high availability in professional hosting and VPS environments built for heavy load.

Today the idea of a cluster is at the heart of container orchestration — Kubernetes manages a cluster of nodes onto which it automatically schedules and scales applications, shifting workloads if any machine fails.