Individual nodes are connected using a very-high-speed computer network and overseen through a centralised managing computer responsible for allocating tasks to nodes
Cluster setup
1. Master computer
2. Network switch
3. Nodes
4. Disk storage
5. Internet connection
Cluster computer
Small cluster computer at the Open University
Cluster nodes
32 core Intel Xeon processor
512 GB of RAM
40 Gbps network connection
Clusters
Relatively cheap and reliable way of processing large amounts of data
Cluster nodes
Relatively cheap
Can be built from consumer PCs or video game consoles
Clusters
More nodes can be added or old nodes replaced as demand increases
Highly resilient to failure of one or more nodes
Grid computers
Closely related to clusters, but made up of many different types of computer all working on the same task, and not necessarily belonging to the same computer network
Distributed computing
Data processing is distributed amongst the nodes, with every node simultaneously working on a tiny part of the entire problem
Parallel processing
A form of distributed computing where the task is distributed amongst many nodes that share memory, unlike cluster computing where each node has its own memory and communicates via a network
