It sometimes feels like The Internet of Things (or IoT) has been around for a very long time and surely, we already have it. For example, this author worked with a well-known high street baker back in the 2000s and they already had internet connected weighing scales then which allowed them to dynamically modify the price of various goods from a central location. However, IoT is just about being connected to the internet. In this blog we will take a look at what is meant by IoT, what an IoT environment needs to provide, whether existing technologies fit into this model and who the vendors of IoT systems are.
What is the Internet of Things?
The Internet of Things is one of those ...things that everyone knows what it is but can't necessarily give a definitive definition or. Some might say it's the internet with lots and lots of things connected to it, or that it's the internet connection of many different devices to services for those devices or perhaps its things that talk to one another via the internet!
According to Wikipedia the Internet of Things (or IoT) “describes devices with sensors, processing ability, software and other technologies that connect and exchange data with other devices and systems over the internet or other communication networks”
So, there are a number of key things here, the ‘things’ are devices which posses processing ability, software on board, typically some form of sensor or sensors to allow them to interact with the world outside the device. These devices connect to other devices and services over the internet. These connections allow them to exchange data with those devices and services.
These devices are perhaps most often associated with the ‘smart home’ concept with appliance such as home security systems, cameras, televisions, entertainment systems, domestic appliances (such as fridges and freezers), heating systems etc. all being part of the things that are connected.
However, today these systems go much further than the home and may encompass, your car, your phone, wearable technology, connected health etc. They may support the care of the elderly via voice control of appliances or remote monitor of elderly relatives, etc.
However, the IoT is not limited to the domestic or personal world, as the so-called Internet of Medical Things (or IoMT) indicates. In IoMT the concept of the basic IoT is applied to medical and health related purposes including data collection, monitoring and analysis. The Industrial Internet of Things (or IIoT) does the same thing for industry, allow industrial devices to acquire, monitor and analyse data from devices, systems, technology, locations and indeed people.
What is needed for an IoT Environment?
Well one thing is for sure, the things, aka devices, need to be connected either together and / or to services. These connections need to allow for data to be exchanged and almost certainly will need to be secure. As of course the name suggests with respect to IoT, that connection is expected to be handled over the internet using internet protocols. However, it’s not just as simple as sticking a device on the internet and expecting it all to work, let’s have a look at what is required at a high level to run an IoTs environment.
Device Registration and Management. It is necessary to have some way to configure, register and provision new devices with the IoT environment. Once a device has been registered it may need to be monitor and manage that device to ensure that it is still available and when it appears the device is no longer responding it may be necessary to trigger some action to allow for investigation of that device. Whether this is automated or involves human intervention depends on the application and the environment.
Remote Control. In many cases it is necessary to be able to modify, reconfigure, restart or otherwise control the functionality of the devices within an IoT environment from a remote or centralised location.
Remote Updates. It is usually considered desirable to be able to update the software deployed onto a device within an IoT system to allow for bug fixes, security patches or any other required updates.
Device Decommissioning. As well as set up new devices, updating existing devices it is often necessary to be able to decommission existing devices.
Data Analysis. Typically, IoT devices generate data that needs to be communicated to other devices and / or services that will analyse that data. This analysis might be bespoke or standard reporting style analysis, however the IoT environment needs to be able to support such analysis.
Device Numbers. The IoT concept is not aiming to allow 10s or even 100s of devices ot be interconnected, here we are talking billions and billions of devices with trillions or more messages being exchanged.
Security. Security is almost always (probably actually always) extremely important within an IoT environment. Do you want others taking control of your home, your car, your pacemaker etc. for example. There is therefore a need for multi-layered security (aka defence in depth) systems. This ensures that there are multiple barriers in the security system that would have to be hacked before anyone can access sensitive data, systems or devices.
AI. The subject of Ai integration comes up almost everywhere nowadays and the IoT is no exception. Many IoT environments are rushing to integrate AI in their environments and to stress to prospective clients how AI enhances their experience. In many cases AI is used to help monitor, report and configure the devices within a system.
What IoT Environments are there?
There are numerous IoT environments available. Some of the most popular choices come from the usual suspects. For example, you have Amazon Web Services (AWS) IoT Core, Microsoft Azure IoT Hub and Google Cloud IoT Core. As you might expect these systems often integrate closely with the vendor’s own cloud solutions, for example, Microsoft’s offering is gaining popularity in manufacturing due to its integration with the Azure cloud.
Other notable platforms include IBM Watson IoT and the Cisco IoT Cloud. However, open-source solutions should not be ignored. For example, ThingsBoard and OpenRemote are both open-source platforms that support IoT type architectures either hosted in the cloud or via on-premisses servers.
IoT and Edge Computing
The cloud is not the only supporting technology clamouring for attention within the IoT world. Edge Computing is gaining increasing importance within IoT systems. This is particularly true for applications requiring real-time data processing and analysis. This is because Edge Computing brings processing power closer to the devices generating the data; enabling faster processing of the data with lower latency times.
Containerization
Within the world of computer science containerization allows software to run within a contained and controlled environment, often insulated from the realities of the underlying hardware being used. Within the contact of IoT this is a potentially very useful and powerful abstraction as well as actual infrastructure.
A container allows a developer to package a piece of software, along with all its dependencies into a ‘container’ and for this container to be deployed to a device within an IoT applications. The container can then be used to ‘spin up’ the device with the software running on it.
There are several different container solutions available which can be used with IoT environments. However, a key aspect of the IoT world is the need to be able to orchestrate and manage these containers. This is often when either the IoT environment comes in or a third-party orchestration tools is used.
Kubernetes
Kubernetes (aka K8s) is a container orchestration tool that enables the scaling of distributed applications by containers across multiple servers and resources. It is an extremely widely used tool within the cloud-native, service (and micro service) oriented architecture style of applications. So why isn't it so widely used within the IoT world?
Kubernetes is a very powerful and feature rich environment that is often too large, complex, resource intensive for IoT systems. This is because the IoT type of application involves billions of often very small, lower capacity, low power, resource constrained devices that are in many cases very widely distributed. Or to put it another way, Kubernetes is a sledgehammer to crack very many little nuts – its simply overkill.
Other challenges for a tool such as Kubernetes is that IoT devices often operate in environments with limited, often intermittent or unreliable network connections, that can disrupt Kubernetes operations. The shear volume and geographically distributed nature of the devices in an IoT application can be make it difficult for Kubernetes to manage. In addition, given that the devices often have limited power / battery life resource intensive systems such as Kubernetes are too ‘expensive’ to run.
Also think about the way in which Kubernetes might push updates out to devices, that might be fine for a cloud-based travel application, but do you really want it updating the software in your car, in real time, while you are driving it! probably not!
Therefore, for IoT environments lighter weight alternatives tend to be used. These may be based on Kubernetes such as K3s or specialised for IoT systems such as balena or Akraino.
Summary
The Internet of Things is becoming a reality and is something which is touching on our everyday lives. However, it is its own subject, using its own infrastructure and with its own challenges.
Would you like to know more?
If you found this article interesting you might be interested in our collection of courses suitable for IoT developers