Fog computing of Fog computing

Concept

The term Fog Computing ("fog computing") was introduced into circulation by the vice president of Cisco company Flavio Bonomi in 2011. He offered the concept of Fog Computing by analogy with Cloud Computing as the clouds expansion to network edges. Technology, the concept of Fog Computing is closely connected with the distributed (cloud) data centers in which servers of data centers can be located in many locations, up to network edge. Data centers can be small (container, modular or mobile execution), being actually "carrying out" of large data centers. Thus, distinctive feature of Fog Computing - nearness to end users and support of their mobility.

Development of Internet of Things (IoT, Internet of Things) demanded support of mobility of IoT devices for different locations with a geolocation and with a small delay on data processing. Therefore the new platform for the satisfaction of such requirements which received the name Fog computing - "fog computing" was offered. Its main feature is data processing in close proximity to sources of their receiving, without the need for their transfer to large data centers only in order that them there to process and transfer back results.

Thus, there is clear an origin of the term "fog computing": when the dense cloud falls to the Earth's surface (by network edge), we see fog.

Standardization

In 2015 for the purpose of development of uniform approaches to Fog implementation by the companies ARM Cisco Dell Intel, Microsoft and also the Princeton University (Princeton University), in USA created consortium OpenFog Consortium. Further, the consortium OpenFog included many other companies (General Electric Hitachi ZTE and dr) and also the universities, for example, of ShanghaiTech University. By 2018 in OpenFog there are more than 50 members. The consortium OpenFog develops standard architecture^[1] (OpenFog Reference Architecture).

Implementation and architecture

Both Cloud, and Fog Computing, use similar IT resources: computing devices (servers and processors of computers of users), switching nodes of network and storage system. However, expansion of a cloud to network edges does not come down only to scaling of this cloud. Technical implementation and also a range of applications of Fog, can differ from Cloud considerably. Fog is intended, generally for applications and services which badly work in architecture of Cloud Computing, or cannot work in it at all. Generally it is area of Internet of Things which accruing development cannot be completely kept only by means of^[2]. Development of IoT was faced the need of filtering and preprocessing of data before sending to a cloud. Generally it following applications:

• The applications requiring low and predictable transfer lag of information on network, for example, game applications or video conferences.
• Applications for transport, such, as: unmanned vehicles, high-speed trains, intelligent transport systems, etc.
• The applications requiring local data processing in real time such, as: intelligent systems of power supply (Smart Grid), intelligent transport systems (ITS), geophysical investigation of a subsoil, control of pipelines, sensor networks of monitoring of the environment and so forth.

Fog is not an alternative for Cloud. On the contrary, Fog fruitfully interacts with Cloud, especially in administration and analytics of data, and such interaction generates a new application class.

The architecture of Fog Computing represents certain "layer" on border between a cloud and devices of Internet of Things with sensors and also mobile devices of users.

Architecture of Fog Computing network

The main architectural differences of Fog from Cloud:

• Ensuring quality of services (QoS, Quality of Service) that requires dynamic adaptation of annexes to network condition.
• Tracking of location (Location Awareness) to maintain stability of operation of application in the conditions of terminal mobility.
• Tracking of the context information (Context Awareness), i.e. capability to detect existence of available resources nearby to involve them in operation of application, with a possibility of horizontal interaction.

In architecture of Fog network points (Fog Sites) located closer to cloud data centers have bigger computing power and бóльшим amount of data in storage systems. The network points located closer to sensors of Internet of Things and mobile devices have bigger interactivity and a fast response. Distinctive feature of Fog is that the user's devices, such as personal computers, home gateways, set-top boxes and mobile devices can act as network point. That the device of the user could work as a node of Fog network, the user should give to the telecom operator the corresponding permission to use of computing power of the gadget in the background, in exchange for different privileges from the operator.

Standard application of Fog Computing

Distributed computing and Big Data in IoT

Standard application of Fog Computing

Platforms for a fog computing

Some scenarios of use

Scenarios of use Fog Computing can be much, and development of adjacent technologies will allow all new scenarios. According to Ovum, Fog Computing is the main environment for Internet of Things^[3].

Trends of Internet of Things and Fog Computing.

Autonomous systems of control of transport (ADS, Autonomous Driving System)

ADS use different multimode sensors, technologies of computer vision and the analysis of images, satellite and network positioning on cards and predictive analytics based on which ADS helps to manage to the driver or manages independently moving vehicle. In such applications high high-speed performance therefore the Fog-node with elements of artificial intelligence needs to be placed directly in the vehicle is required.^[4].

Components of Fog-nodes in ADS

Fog-systems in electronic health care (eHealth)

Fog-systems in medicine are applied when it is necessary to make the operational analysis of data retrieveds from wearable the patient of sensors and to take emergency actions according to the plan of treatment^[5].

For example, Fog-technologies are already applied to status monitoring of patients with diabetes and automatic introduction ^[6]. The sensor on a body of the patient defines critical value of content of sugar in blood, and through Fog-network issues a signal on accomplishment of an injection by means of the micro syringe which is also located on the patient's body. Thus, the patient gets rid of need to constantly perform measurements and to inject most.

Fog-projects of cloud providers

In 2016 three largest providers of cloud platforms - Amazon, Google and Microsoft - began several projects of use Fog Computing in the ecosystems of IoT in which the so-called "bezserverny architecture" (serverless architecture) is applied.

The Bezserverny architecture allows to execute the source code of thousands and millions of users (in particular, fog-devices) in computing environment, without caring for scaling of resources.

Microsoft

The Microsoft company announced support of the Azure functions (Azure Functions) in the platform of development of SDK (Software Development Kit). The Azure functions were entered into families of the cloud products with bezserverny architecture (Serverless Architecture) developed in Microsoft in the beginning.

Amazon

The Amazon company developed the Greengrass platform with support of so-called Lambda-functions (bezserverny architecture) in IoT devices in interaction with a cloud platform of AWS. Greengrass is a container of execution of a software module which can be started directly on the Fog-device, but not on the server in data center. Devices can communicate with Greengrass among themselves regardless of existence of the external Internet, i.e. is horizontal between Fog-devices by means of different radio-internet protocols of things.

Google

Google submitted the platform for Internet of Things of Android Things with support of microcomputers of Intel Edison and Joule 570x, NXP Pico i.MX6UL and Argon i.MX6UL and also Raspberry Pi 3. Fog-applications are developed on the Android Studio platform for any of these devices. Android Things also provides integration with Google Play and all ecosystem of Android at which 90% of smartphones in the world work now. Thus, the Android Things system gives the chance to any Android smartphone or the tablet to work as a Fog-node.

These projects show a trend of "kommoditization" of IoT devices, i.e. design and their creation based on the public Fog Computing elements.

In Russia IoT and Fog Computing technologies are used, for example, in the solutions "intellectual pit" of the Russian company "VIST Mining Technology" (the word "mining" in the name of the Russian company is used in the original value – mineral extraction).

Scheme of the solution "intellectual pit" of the Russian company VIST Mining Technology

Advantages and forecasts

Fog Computing is a new step of development of cloud computing which reduces the delays arising at data transmission in the central cloud and provides new opportunities of creation of intelligent devices of the Internet ^[7]

Advantage of a fog computing is data reduction, told to a cloud that reduces requirements to network transmission capacity, increases data processing rate and reduces delays in decision making. A fog computing solves a number of the most widespread problems among which:

• high delay in network;
• the difficulties connected with mobility of termination points;
• communication loss;
• high cost of bandwidth;
• unforeseen network jams;
• big geographical distribution of systems and clients.

Global market of Fog-systems is estimated at $18 billion by 2022^[8].

Have the most high potential of development of Fog computing technology in the following industries: power, municipal services, and transport, agriculture, trade and also health care and industrial production.

The energy sector and municipal services represent the greatest market for the Fog computing systems, with a potential of growth up to $3.84 billion by 2022.

The sector of transport – the second most important potential market for Fog computing with a potential of growth up to $3.29 billion by 2022

The industry of medicine represents the third largest Fog computing market which amount is estimated at $2.74 billion by 2022.

Development in Russia

2016: The Kremlin charged to prepare infrastructure of a fog computing

On July 1, 2016 it became known of what the Kremlin charged to the Ministry of Telecom and Mass Communications, to Minpromtorg  and also  other departments to prepare infrastructure of a fog computing.

According to Kommersant, the Presidential Administration directed in the Ministry of Telecom and Mass Communications, Minpromtorg, Rostelecom and the Agency of Strategic Initiatives (ASI) the letter with the signature of the president Vladimir Putin on development of a fog computing in Russia.

Putin charged development of a fog computing in Russia

It is entrusted to work implementation of a fog computing in economy of Russia and also creation of the hardware and software systems necessary for work of infrastructure of a fog computing — reported the edition of people, close to the Kremlin.

Other source in one of  the profile ministries in  a conversation with  the newspaper specified that  results of accomplishment of order need to be provided in  October, 2016. It is noted that the initiative of work in this sphere proceeds from Rostelecom and ASI whose idea was supported in Presidential Administration.

Kommersant reports that Minpromtorg appealed to several Russian technology companies to provide project appraisal. In particular, the ministry sent the corresponding request to T-Platforms company which is engaged in development of ADP equipment.^[9]

Notes