Fog computing performed on servers located outside the cloud.  It allows data points to be gathered closer to the data source and compiled into a daily report that is sent to the data center. In other words, fog computing helps to extend cloud computing by transmitting computations onto the network edge. In this post, we describe fog computing basics and discuss possible applications.

Fog computing basics

Continuously rising IoT connectivity requires low latency, mobility, and location-awareness, therefore, fog computing could be an option addressing such challenges efficiently. Fog computing enables to move the necessary processing and calculations closer to the data source, eliminating the necessity of remote data processing, and improving data transport performance.

Edge computing is where a particular IoT or some other device combines the readings from various sensors to a single data point that gets sent every particular time into the fog.  In its turn, fog allows those data points to be gathered closer to the data source and compiled into a daily report that is sent to the data center. In other words, fog computing helps to extend cloud computing by transmitting computations onto the network edge.

Deploying so-called fog nodes (routers, various controllers, cameras, and devices) in some target areas or even within vehicles, makes it possible to analyze datasets generated by IoT devices and to avoid data transfer to the cloud, providing decentralized local access.

Fog architecture

Let’s briefly go through a six-layer fog computing architecture. The physical and virtualization layer includes various nodes virtual sensors and virtual sensor networks. Distributed virtual sensors collecting data and send it to upper layers for processing by means of gateways. The monitoring layer is responsible for activities monitoring, with all the nodes tasks monitored within it, alongside with a status of services and applications. Data management tasks (analysis, filtering, etc.) implemented within the pre-processing layer and acquired data further stored in temporary storage layer.

The security layer on its turn is responsible for encryption/decryption of obtained data and provides opportunities for integrity measurements. As a result, the final transport-layer supports pre-processed data uploading to the cloud. Data uploading appears to happen portion by portion to enable efficient power usage.

Implementation and applications

Fog computing could be treated as an extension of cloud computing that more actively interacts with IoT devices and relevant data. Autonomous vehicles is rapidly evolving trend where fog computing could be employed to provide further progress in “hands-free” driving via enabling highly efficient real-time interaction. In addition, reduced latency time could potentially improve overall efficiency, collision prevention, further advance fleet management, allow smart traffic lights and various wireless sensor/actuator networks. Lets further go through an example of a gateway based fog computing architecture for wireless sensors and actuator networks, following [W. Lee et al, 2016].

The proposed architecture for WSANs is shown in the figure above. The suggested fog computing consists of the sets of gateway and microserver. The gateway can be a conventional or smart gateway with high performance and diverse interfaces. Gateway and microserver are connected by the Ethernet interface. Interfaces between gateway nodes are wired and wireless interfaces such as 3G, LTE, and Ethernet. It is essentially an event-driven virtualization model, providing virtual events from networked objects and sharing them with various applications.

Fog computing extending cloud computing opportunities and helps to move the necessary processing and calculations closer to the network edge. It provides increased latency, flexibility, real-time analytics, reduced performance requirements, and opens new ways to utilize a broad range of IoT devices in various applications.

References