From Grid Computing to Cloud and Beyond

Centralised Computing

All resources in one central machine.

Cluster Computing

Multiple systems work together to work as a single machine.

Grid Computing

Distributed resources across organisations are coordinated to achieve a common goal.

Working of Grid Computing

Types of Grids

Computational Grids

• Network of high computational servers.

Scavenging Grids

• Network of Desktop computers.

Data Grids

• Access, modify or transfer huge amounts of data

Advantages of Grid Computing

• Improved resource utilisation
• General performance increase (parallel processing)
• Easier collaboration
• Increased robustness

Disadvantages of Grid Computing

• Difficult to set up initially
• Huge administrative overhead
• Resources are not guaranteed
• Not suitable for dynamic workload

Utility Computing

• Pay-as-you-use model

Cloud Computing

• Grid Computing + Utility Computing
• Resources are provided by Cloud Service Provider
• Pay-as-you-use model

Edge Computing

• Computation happens at the edge device
• eg. Self-driving cars

Advantages

• Low latency
• Faster decisions
• Privacy

Disadvantages

• Lack of global view
• Low computational power

Fog Computing

• Decentralised architecture that places fog nodes between the edge devices and servers to handle intermediate computation.

eg. Traffic light system.

The sensors (edge devices) fetch the data and the fog nodes processes the information for immediate decisions. The cloud layer may fetch the summary from each fog node to optimise city’s long-term traffic control.