- Model question
Grid, cluster and cloud systems
Short answer
Cluster computing uses homogeneous nodes on a fast LAN under one administration. Grid computing federates heterogeneous resources from multiple administrative domains using middleware. Cloud computing provides virtualized resources as IaaS, PaaS, or SaaS over the Internet with elastic, pay-per-use scaling.
Explain the different types of distributed systems: grid, cluster, cloud.
These are the main types of distributed computing systems, used for high-performance computing.
Cluster computing
A set of similar computers (nodes), connected by a high-speed LAN, working together as a single system under one administration, with each node usually running the same OS.
- Nodes are homogeneous and close together.
- A master node distributes the jobs to compute nodes.
- Uses: high-performance computing, web server farms, load balancing, high availability.
- Example: a Beowulf cluster.
Grid computing
A federation of heterogeneous computer systems from different administrative domains and sites, combined by software (middleware, for example Globus) to work on large tasks.
- Nodes differ in hardware, OS, network and security policies.
- Resources are shared by virtual organisations.
- Uses: scientific problems, such as particle physics (CERN LHC grid), SETI@home.
- Layers: fabric, connectivity, resource, collective, application.
Cloud computing
On-demand delivery of computing resources (servers, storage, applications) as a service over the Internet, with pay-per-use billing and elastic scaling.
- Service models: IaaS (virtual machines, AWS EC2), PaaS (Google App Engine) and SaaS (Gmail).
- Deployment models: public, private, hybrid, community.
- Resources are virtualised and shared by many tenants.
| Point | Cluster | Grid | Cloud |
|---|---|---|---|
| Nodes | Homogeneous, one site | Heterogeneous, many sites | Virtualised data centres |
| Administration | Single | Multiple | Provider-managed |
| Connection | Fast LAN | Wide-area network | Internet |
| Billing | Own hardware | Resource sharing | Pay per use |
| Scaling | Fixed, manual | By joining sites | Elastic, automatic |
✓ Answer checked by Claude Opus 5.5 (high effort)
Related questions
- Why NOS preferred over DOS · Distributed System, chapter 3 · asked 2×
- Heterogeneous model; DOS realization · Distributed System, chapter 3
- Cloud and grid computing concepts · Internet and Intranet, chapter 7 · asked 2×
- pipe(); semaphore in distributed systems · Operating System, chapter 3
- Distributed system and Hadoop · Information System, chapter 8
- Lumped vs distributed analysis, applications, protocols · RF and Microwave Engineering, chapter 1