Internet Of Things

Understanding IoT Architecture in a simple Way

Building an Architecture means by considering your requirements building a blueprint that’s the general expectation however IoT was not “designed”, it just “Happened”. In the past Engineers from different organizations with different requirements adapted multiple sensor type, multiple protocols as a result of this IoT architecture is a bit complex

IoT Architectural challenges

While designing IoT Architecture generally following challenges are faced.

• Large Scale: If you have millions of clients in a single network then IPV4 is not useful you have to use IPV6
• Security: In a network sensor are exposed to the world also data travels through public network. So, security is an important concern. You may use encryption for that.
• Constrained Devices: Power consumption is an important aspect in IoT architecture. So IoT devices which we are using in network used by taking some characteristics into account like size, battery, storage, CPU power
• Large Volume: If you have millions of devices then you have millions of packets. So while designing application you have to take this huge data in account.
• Real time data exchange: while designing your IoT application the information must be exchanged in real time in the network.

IoT Architecture Models‍

There are lot of Architectural models because IoT architecture is application specific. IoT architecture may differ with different requirements. But in this blog we are going to take a quick look of two important architectures as follows.

• ESTI M2M standardized Architecture: This architecture is developed by European Standard Telecommunication Institute. This architecture is used as a reference to describe the communication to the IoT world. This consists of three Domains or layers.
• M2MDevice Domain: M2Mmeans machine to machine device domain. In this domain you will find sensors communicating with each other and with the network. To communicate with network, they need to have a gateway outside of the local network inside a backbone they will bring data to an application. At this layer you will find some protocols like IEEE 802.15.4, IEEE 1901PLC, IEEE 802.11ah.
• Network Domain: This is the second domain of architecture. Network domain is the connectivity between local area and the some application somewhere in the cloud. We can use satellite connectivity, wired connectivity or wireless connectivity. You can find different kinds of protocols here like LTE, WiMAX, MPLS. Basically by using this domain we can establish connectivity between Device domain and Application domain.
• Application Domain: Application domain is the domain where application is created by collecting data from devices of the devices domain
• IoT World Forum (IoTWF) Standardized Architecture : In2014 IoTWF architectural committee led by tech giants like Cisco, IBM and others a seven-layer IoT architectural reference model. It looks like networking OSI layer. This architecture describes perspective in a simplified manner. It is different from ESTI M2M. This architecture includes edge computing, data storage and access.
• The seven layers are lucidly explained further

1. Physical devices and controllers: This are the first layer of IoTWF Standardized architecture. This layer consists of sensors or actuators devices, machines, smart devices which may be present in your home or streets or anywhere.
2. Connectivity Layer:  The objects in physical layer needs to connect somewhere so that’s why we need connectivity layer. This layer provides two layered connectivity. This layer is for just local communication so that we can take data out of sensors and other devices in physical layer and store it somewhere locally. This is the local connecting point where we don’t need IP routing yet.  
3. Edge Computing Layer: This may be the place near your sensors where you can process your data which we have collected in connectivity layer. We can evaluate and data for processing at higher levels. We can also access data for alerting, notification, or other actions. In this layer we just process our data locally. Then data may be sent to somewhere probably to central application.
4. Data Accumulation (Storage) Layer: At this layer we can store the data which we got from Edge Computing Layer. The data can be processed or raw. We can look into data and make intelligence out of it by using machine learning. To make data available at the central location is the key point of IoTWF Standardized architecture.
5. Data Abstraction Layer: At this layer we reuse the data and abstract it. By using some algorithms we can gather some intelligence from numbers like something is too hot or too cold.
6. Application Layer: At this layer we can find different IoT applications depending upon their purpose.
7. Collaboration and Processes Layer: At this layer we can find people or minds who are behind the IoT application. Data is used to change the business processes, to send the alarms and for a lot of other things.

A Simplified IoT Architecture

There are some common characteristics of every IoT architecture model i.e. every model is built on layer. There layers are usually independent. Each layer can use multiple protocols. Some protocols can be appeared at several layers also. In every IoT model in the end we need to connect sensors to a network and communicate with the applications. By using these common properties we can derive a simplified model for IoT Architectures. In most of the IoT architecture you can see that device layer, network layer, application layer is present.

In the most of the architectural models the data management has been parallelly done with layer-to-layer communication. At the sensors layers some kind of data processing can be done we can call it mist layer or edge layer. We can have cloud to whole manage data at application layer. At network layer we have fog layer to manage some intermediate data. These IoT functional layers and database layers works parallelly. We can see these two parallel domains of three layers almost everywhere. And at the same time with these two domains security is everywhere. If you consider this simplified model you can observe every IoT model and understand it very easily.

Source: https://medium.com

IoT Connectivity Landscape

IoT communication model is based on standard IT communication model with some modifications. IoT communication model is a bit more complex than standard IT model. In standard IT model endpoints i.e., laptops, tablets connect with data center or cloud by using network where some kind of application might be running. IT connectivity architecture is a three-layer connectivity architecture.

The IoT connectivity model which based on standard IT connectivity model is of four layers. The first layer is of devices i.e. sensors or smart devices. But the second layer is fog layer which is before network layer. The third layer is of Network Layer. In huge IoT networks we can’t use IPV4 we have to use IPV6 there as there are millions of devices connected to each other and to the network. The last layer is of data center or cloud. The IoT communication model based upon IT communication model is shown in following image.

Source: http://www.ijste.org

Range, Throughput and Scale

Sensors characteristics: They are based upon IoT requirements. In the past some IoT requirements used existing protocols while as others created different protocols as per their IoT needs. So, there are various type of protocols we use in IoT. We need protocols for transfer information from sensors to the application. So, it is important to choose right protocols based upon your IoT requirements. For example if you have powered your sensor by using cable then there are no limitations on you how much you are sending and how often you are sending. But if you have powered your sensor by using battery then there are limitations on how much data you are sending and how often you are sending. If you are sending data often and you are sending more data so it will drain your battery early. Depending upon these things you need to choose your protocols wisely.

Mobility: If you have powered your device by using cable then it is not going to be mobile. If you have not powered your sensors by using cable then it is mobile i.e. it can be connected to an object which as a power source not far from sensors. If your device doesn’t have power source nearby then you need to decide how much power your device is going to consume and for how much time it is going to send information. Mobility means power and it also means. If you want to transmit data over miles then it will consume more power.

Low and High Reporting Frequency: It means you need to send the data how often i.e., thousands of times per second or once in a month or once in a day etc.

Simple or Rich Data: Less power is required to transmit simple data with one parameter where as more power is required to transmit complex data with thousands of parameters.

Transmission Range: It means how long you want to transmit data. If you want to transmit data over miles then you need robust technology for that. Longer distance travel means more chances to lose data. You must take this point in account while creating a IoT application.

Density Per Call:  if we are using few devices then there are less chances of collision of data but if there are huge number of devices in the network then more chances of data collision.

Range and Transmission Technologies

There are different technologies available to transmit the data. Every technology differs in terms of range of transmission. The following technologies are used to transmit over 0 to 5km

WPAN: Wireless Personal Area Network consists of technologies like Bluetooth, MI WI, ANT+

WHAN: Wireless Home Area Network consists of ZigBee, Z-Wave, Thread and many more technologies.

WFAN: Wireless Field Area Network consists of Wireless Hart, ISA 100.11a

WLAN: Wireless Local Area Network consists of technologies like 802.11a/b/g, 802.11ah, 802.11p etc.

WNAN: Wireless Neighborhood Area Network it consists of Wi-Sun, ZigBee NAN and many other technologies.

The following technologies are used to transmit data in the range of less than 100km

WWAN: Wireless Wide Area Network consists of Cellular like WCDMA, WiMAX and licenced 3G, 4G services.

LPWA: Low Power Wide Area consists of SIGFOX, LoRA, Telensa, Onramp, Positive Train Control and many other technologies. It has longer range but due to low power we can’t send much data.          

Refer the following figure for betterunderstanding of IoT Transmission technologies.

Source: https://html.scirp.org

Contact us for expertise training in ROBOTICS