What is cluster computing?

Cluster computing is a form of parallel or distributed computing where a group of loosely or tightly coupled computers, called nodes, work together to perform a common task. The nodes share resources such as CPU, memory, and storage, and communicate over a network. Clusters can be homogeneous (identical hardware) or heterogeneous. They are often used to increase computational throughput, provide fault tolerance, and improve scalability. Key components include a job scheduler (e.g., SLURM, PBS), a shared file system (e.g., Lustre, GlusterFS), and a communication layer (MPI, gRPC). Examples of cluster workloads include scientific simulations, big‑data analytics, and machine‑learning training. Benefits include cost‑effective scaling (adding commodity servers), high availability (failover to healthy nodes), and load balancing. Design considerations involve network topology, data locality, and consistency models. In practice, a cluster might run a Hadoop YARN cluster for batch processing or a Kubernetes cluster for container orchestration, each leveraging the underlying compute nodes to deliver scalable services.