Edge computing represents the data ingestion, processing and information communication of any endpoint device at the edge of a network, such as a microcontroller unit, PC, smartphone or gateway device.

Fog computing represents the movement of traditional cloud services, such as networking, processing and application enablement to the edge of the network. Fog computing devices are typically smart gateway units connected to a local area network.

A loose representation of the architecture for these concepts is shown in the following figure.

Impact of Edge Computing

• Latency: that is to say, the time taken between data being recorded, analyzed and actioned can be considerably reduced. If edge devices are able to identify relevant data, and process it meaningfully, then that data must no longer be sent to a potentially far-off cloud server. As such, real-time or near-real-time applications can be enabled.
  

• Bandwidth is reduced, owing to less reliance on cloud computing services. As the IoT scales up, this reduction in bandwidth will reduce the likelihood of network bottlenecks.
  

• Reliability is increased, as deployments become less dependent on Internet connectivity uptime.
  

• Data security can be increased, by virtue of the ability to better anonymize and group sensitive data.
  

• Automation potential is increased via an increase of on-device machine learning capability closer to the data source.