(Source: https://unsplash.com/s/photos/internet-of-thing)
Web technologies enabled the vision of realizing a network of computers at the global level. Similarly, the idea of the Internet of Things (IoT) is to build a network of all devices or things globally. It is possible that home appliances like electric bulb, microwave oven, and air conditioners to communicate with each other as computers communicate when they are connected through the internet. Similarly, it is also possible to connect vehicles such as cars, automobiles, bicycles, and trains, transferring their current information about location, fuel consumption, and speed. In the manufacturing sector, industrial machines and automation plant can be connected through IoT to schedule activities and increase a manufacturing plant's productivity.
There are various definitions of IoT, but one widely accepted definition of IoT is quoted here:
The worldwide network of interconnected objects uniquely addressable based on standard communication protocols.
Also equally acceptable and the broad definition of IoT in terms of what is expected from the IoT infrastructure is given below:
Interconnection of sensing and actuating devices providing the ability to share information across platforms through a unified framework, developing a common operating picture for enabling innovative applications. This is achieved by seamless large scale sensing, data analytics and information representation using cutting edge ubiquitous sensing and cloud computing.
Functional Elements of IoT
The basic functions performed in an IoT system include:
* Identification, The number of possible devices that can be connected to IoT is huge. Identifying an individual device out of these large number of devices is essential for communication.
Often the IoT systems differentiate between an identifier associated with a device and its address used for communication purposes. The identifier is a name associated with a sensor or an actuator, or any device such as T1 may be an identifier associated with a temperature sensor. The address of a device is used for communication and data transfer purposes, and it must be unique within a network. Assigning a unique identifier to a considerable number of devices in IoT is a difficult task. Current addresses that follow IP version 4 (IPv4) to identify computers on the Internet has limitations because of their insufficient address space. In addition to IPv4, addressing methods used in IoT include IP version 6 (IPv6), Electronic Product Codes (EPC), ubiquitous Code (uCode), RFID and 6LoWPAN. These are briefly described below.
* Radio Frequency Identification (RFID) plays a dual role in IoT. It can be used for identification and communication purposes. The RFID tags associated with an object responds with an identification number when queried by an RFID reader.
* EPC The EPCs are Unique Resource Identifiers (URI) similar to Unique resource Locators (URL) used on the web. The EPCglobal standard defines the data structure and format for URI or uCode. It is an open standard used to referer to physical objects in a business organization.
* uCode is an addressing scheme that uses 128-bit long and extensible in the chunks of the 128-bit numbering system. The code is available in various media such as printed code, QR code, and RFID passive tags.
* 6LoWPAN:}This addressing scheme uses IPv6 to identify IoT devices using IP addresses uniquely. The acronym 6LoWPAN stands for IPv6 over Low-Power Wireless Personal Area Networks.
The IoT system may use any one or more addressing system to address a device in the network uniquely. The software provides interoperability among these addressing schemes.
* Sensing The objective of sensing is to collect the physical world's data and transfer it remote machine or a cloud server. The data is typically collected through sensors, actuators and wearable devices. A remote machine processes the data. Many IoT products are equipped with single board computers, sensors for temperature, humidity, level detector, and pressure change, convert the collected signal into digital format and communicate information using TCP/IP. Few examples of such products are Arduino and Raspberry Pi.
* Communication The IoT devices work in low-power, low-bandwidth, lossy wireless communication medium. Any communication protocols being adopted for IoT need to consider these constraints. Some of the communication protocols that are commonly used for IoT are described below:
* RFID This protocol is used for communication between RFID tagged objects and an RFID reader. The RFID-based systems come in two different variants called passive and active RFID. A passive RFID does not need any battery to supply power, and they derive energy from an RFID signal transmitted by an RFID reader. A battery-operated RFID is known as an active RFID. Physically RFID is available in the form of adhesive strips in different sizes. An RFID reader initiates communication by transmitting a query signal to which an RFID tag responds. Thus RFID protocol is quite helpful to monitor devices in a network within a range of 10 meters.
* Near Field Communication (NFC) This form of communication protocol is useful to transfer data in short-range, i.e. up to 10cm. This protocol is primarily used for contactless authentication and payment mechanism. This protocol achieves communication is between active/passive tags and a reader.
* Wireless Sensor Networks (WSN) Wireless communication protocol is preferred to design IoT systems. A wireless sensor network consists of a sufficiently large number of sensors nodes communicating to a node called sink nodes. These sensors acquire real-world physical parameters such as temperature, humidity and other parameters. These WSN use the IEEE 802.15.4 standard defined for low-power, low bit-rate communications in wireless personal area networks (WPAN). A WSN can operate in the range of 100m, and it is also possible to integrate with RFID based systems.
Besides these protocols, some of the commonly used protocols include Bluetooth, Long-Term-Evolution (LTE) and ultra-wideband (UWB).
* Computation, The majority of IoT platforms rely on distributed computing paradigm. Many single-board computers are equipped with necessary hardware such as micro-controllers, memory, wireless communication adaptor, and digital-analogue converters, along with sensors and actuators. This kind of hardware performs local processing of sensed data and transfer them to the remote machine. The real-time operating system is also an essential component of such single board computers. Some of the widely used RTOS in IoT systems include Contiki, TinyOS, LiteOS and Android. The single-board computers offload the collected data to either a remote computer or a cloud platform or a fog node for analysis, storage and visualization purposes.
* Services Many IoT platform provides primary services required to implement higher-level applications. These are often categorized into four types. (i) Identification services: This includes assigning a unique address and converting the identifiers or names of devices into their physical addresses. (ii) Information Aggregation services: They collect and summarize raw sensory measurements that need to be processed and reported to the IoT application. (iii)Collaborative-Aware Services act on top of Information Aggregation Services and use the obtained data to decide and react accordingly. (iv) Ubiquitous Services aim to provide Collaborative-Aware Services anytime they are needed to anyone who needs them anywhere
* Semantics These functions refer to all activities related to providing meaningful information to the users of IoT platforms to make the right decision based on the presented information. It includes extracting knowledge from data, analyzing it, visualizing it, and classifying the collected data into different categories.
A Five Layer Architecture of IoT
The software vendor develops IoT products that realize the above functionalities. Generally, they adopt a five-layer architecture style to implement these functionalities. Some three-layered products also exist, but the 5-layer architecture style is more comprehensive in terms of functionalities; hence, this section reviews the 5-layered architecture for IoT products. The five layers are shown in Figure. These layers are described below.
* Objects This layer includes hardware elements such as sensors, actuators and wearable devices. They collect information from the physical world such as temperature, pressure, vibrations, and locations are some of the parameters they sense usually. This layer is also known as the perception layer.
* Objects abstraction layer This layer provides a uniform way to access any device connected to IoT and bridges the heterogeneity. It transfers the data collected by sensors and actuators to the service management layer over the secured communication channel. This layer typically use protocols such as RFID, LTE, and wireless communication protocols. Also, some of the processing needed for data management and cloud computing is performed at this layer.
* Service Management This layer is responsible for implementing the client-server model of service interaction. So the layer receives and replies to messages, transform these requests to underlying objects and the object abstraction model. The layer is a central element because it integrates IoT over the web.
* Application Layer This layer realizes requests that are initiated by users' applications. For example, responding to query related to temperature/location sensor.
* Business Layer The business layer is responsible for implementing a business process workflow and high-level data analysis and visualization to support decision making processes.
The deployment of IoT platform is raising many system-level quality issues. For example, providing scalability or scalable performance is one such issue that is concerned with providing guaranteed performance when the number of devices increases. Similarly, providing access to any device from anywhere at any time is also encroaching upon people's privacy and threatening their security. IoT is often integrated with other technologies such as blockchain technology, cloud and fog computing to address some of these challenges.
