Referring to the natural composition structure of IoT technology, as well as the layered architecture of the information industry pattern and the business perspective of IoT, the technology matrix of IoT can be divided into six levels. From bottom to top: Element Layer, Device Layer, Terminal and Node Layer, (Information) Resource Aggregation Layer, Platform Service Layer, Application Layer.

The technology matrix is divided into two “domains”: “Edge Domain” and “Cloud Domain”, with the boundaries of the two domains mainly reflected in the Terminal and Node Layer, Resource Aggregation Layer, and Application Layer.

1. Element Layer

The natural phenomena and effects in information science are a collection of various basic information sciences and are the foundation for constructing the device layer (the most basic technical elements of IoT). This includes electricity, electromagnetic induction, microwave principles, circuit theory, etc.

2. Device Layer

The basic components in the information industry are mainly various electronic components, circuit modules, and functional boards, including chips, circuit boards, power supplies, memory, signal input/output devices, sensors, and actuators. In this layer, “software” includes not only circuit design but also programs for embedded development, which are basically fixed before the electronic devices leave the factory.

3. Terminal and Node Layer

Mainly includes two types of physical entities:

A. IoT terminals, including non-smart terminals (also known as “dumb terminals”, generally without data processing capabilities, can only report sensing data through the network or receive control data), smart terminal devices (e.g., industrial robots), and user smart terminals (phones, laptops).

B. Dedicated information processing nodes: network devices, computing devices, storage devices. Although these nodes all have network interfaces, computing, and storage components, they usually provide a certain dedicated information capability. For example: routers in the operator’s IP bearer network (network), blade servers in cloud computing (computing).

Except for dumb terminals, other devices have general or dedicated computing capabilities.

The “software” at this layer mainly consists of various terminal programs, including embedded programs, operating systems, middleware, and general or customized application software.

4. Resource Aggregation Layer

Various information resources converge at this layer. It mainly includes three core information processing resources and capabilities: network (wireless network coverage, IP backbone network, Internet, and interconnection of operator networks), computing (cloud computing, big data, machine learning), and storage (cloud storage, databases).

The “software” at this layer, in addition to various programs running in terminals, also includes various protocols, standards, process specifications, service interfaces, etc. At this layer, the informatization capability has semantic characteristics, possesses a certain openness, and continues to develop towards a more open, flexible, and modular direction.

For example: outside the telecommunications operator system, virtual operators can sign commercial contracts with operators in multiple countries, build their own networks on top of the networks of operators in various countries, and after aggregating the resources of operators from various countries, form a globally covering network, becoming an “international telecommunications operator”. Of course, those specialized network service companies that do not have the funds to build a globally covering network can rent the resources of virtual operators to build globally interoperable SDN networks for some specialized applications.

This method of continuously layering “networks” on top of “networks” reflects the semantic characteristics of information technology.

5. Platform Service Layer

At this layer, not only is information technology semantically defined, but other industry technologies and entities also achieve semanticization (e.g., “digital mapping” in the industrial field—“digital twins”).

In order to support the digital transformation of various industry fields, the various information technologies contained in IoT itself must achieve semanticization. Beyond the field of IoT technology, the resources and capabilities of various industries are also modularly encapsulated here, providing services to the outside world through standardized information interfaces, and application developers and operators can use them as needed.

At the same time, this layer also provides tools for software development and operation, facilitating the use, combination, evaluation, and maintenance of technologies and facilities in various industry systems. In the platform service layer, a dictionary of IoT applications is formed (vocabulary collection), various IoT resources (facilities) are “nouns”, and various information processing functions are “verbs”. Developers can focus their main efforts on constructing “sentences” and “paragraphs” (logical functions of applications) and finally write a complete article (application systems, business systems).

At this layer, the “software” includes the architecture of the IoT service platform, middleware, interface standards, information service specifications, etc., and also integrates some foundational technical services from various industries (for example, the platform can provide weather forecasting services).

Although the platform service layer of IoT makes the design and combination of “IoT technology and various professional technologies” more convenient, the underlying technology is not completely “transparent” to application developers. Developers need to clarify the applicable scope, costs, characteristics, etc. of the underlying key IoT technologies when constructing application logic and utilizing data resources and tools.

Only by deeply understanding the “context” of using the “vocabulary” of IoT can one weave a “rigorous and beautiful” “good article” (IoT application).

At this layer, some already modularized technologies in various industries can provide open services to the outside world through the (programmatic) encapsulation of the information industry, forming a public language (vocabulary) for various industries in IoT.

However, this language is not a language for “people”, but a language tool for “things” to communicate information between devices and systems. As long as one masters the “language” of a certain IoT application, any device can understand the meaning of the application data, use the IoT service of this application, operate remote devices, and obtain feedback.

6. Application Layer

Designing, building, and operating applications for the IoT industry. Developers and operators at the application layer select the required resources and functions (including information technology and industry technology), and construct their own industry applications by invoking the services of the lower layers. In the hierarchical system of IoT technology, the five layers below the application layer can be understood as the “hardware” of the application layer; the “software” is the application logic written by the application developers themselves.

If the application software is hosted on SAAS (Software as a Service), and the underlying system devices are provided by cloud computing service providers, then the application may not see any physical entity “hardware”; instead, it is the modularized and standardized information services and industry functions from the lower layers.

The application layer and platform service layer can also be understood as the “innovation layer” of IoT, a level where application developers can freely play, achieving “combinatorial evolution” through the integration and reuse of technologies from various industries, leading to industrial upgrades.

The two demands of “accessing” and “aggregating” information flow naturally divide the technology of IoT into two “domains”: the “edge domain” where countless devices need to be connected, close to the physical entity site; and the “cloud domain” which is far from real objects but concentrates and aggregates information. At the application level, different functional requirements further promote the separation of the two domains.

The applications and functions of the “edge domain” are more inclined towards real-time feedback operations, while the applications of the “cloud domain” are more inclined towards abstract predictive analysis, massive data retrieval, etc.

In the terminal and node layer, and the resource aggregation layer, the hardware and software contained in the two domains are different. For example, in the edge network, there are various terminals equipped with sensors—used to perceive the physical world; but the devices in the cloud network only have standard signal input/output interfaces.

The edge network will deploy various types of network protocols, including wired and wireless networks, “peer-to-peer network protocols” and “non-peer networks” (Note: “TCP/IP” is a peer-to-peer network protocol, while RFID technology on the wireless side is not a peer-to-peer network protocol); while the “cloud domain” only has peer-to-peer network protocols such as IPv4 and IPv6.

Author: Wang Yiming

Source: IoT Jianghu (iot521)

Network Optimization Freelance Submission Email: [email protected]

Long press the QR code to follow

On the road of communication, let’s walk together!