This is a translation of a blog post originally published on the SATEC website on November 15, 2016.
The term “Fog Computing” comes from the well-known “Cloud” structure, and continues to develop in parallel with the expansion of IoT (Internet of Things) and IoE (Internet of Everything) and the massive deployment of devices connected to the network. If the Cloud Computing concept refers to the distribution and use of services in a network (usually the Internet), Fog is a paradigm that tries to bring data and services as close as possible to the users who need them, this place being called the “edge of the network” and we could say that it is the area between the users’ devices and users’ connections to the network.
Fog Computing, also called Fogging by some authors, is defined as a TIC architecture based on the collaborative use of user devices connected to the network that supports complete functionalities of storage, configuration, management, etc.
Therefore, Fog Computing is an architecture model in which data, services, and applications are concentrated on various devices at the network edge (data can be processed locally on an intelligent device instead of being sent via network). This way, the data transfer through the network, the massively distributed storage, the needs of network management, etc. are drastically reduced in relation to the Cloud model by bringing the services and data used in that Fog environment closer to the network edge.
The Fog architecture is structured from a layer with two separate and well-defined support components called the Data Plane and the Control Plane.
The data plane is responsible for tasks such as:
- Local content and bandwidth management;
- Cache at the network edge;
- Direct client-client communications (such as FlashLinQ, LTE Direct, WiFi, Direct, Air Drop);
- Definition and support of Cloudets (small-scale clouds in mobility and at the network edge).
Additionally, the control plane (less developed so far) is responsible for tasks such as:
- Content Management Over the Top (OTT);
- Fog-RAN or radiofrequency access;
- Client-based heterogeneous network control (HetNets);
- Customer-controlled cloud storage;
- Session management and edge signaling.
In terms of characteristics, the objective of the Fog Computing paradigm is to offer the following:
- Location of data and applications at the network edge and its access with small latencies.
- Distributed computing and storage.
- Deployment with large number of nodes and with very high geographical distribution.
- Mobility support.
- Real-time communications and applications.
- Major wireless access.
- Support and management of a heterogeneous network (devices, data, applications, etc.).
- Interoperability and federation (suppliers).
- Interaction with cloud systems.
Although we are facing an incipient architecture, deployments are already being made to support services and applications in diverse areas such as smart cities, vehicular networks, IoT and M2M systems, Smart Grids, etc. which suggests that its development will continue in the near future.