Click the “blue font” above and select “Set as Favorite“
Key information, D1 time delivery!
The Internet of Things (IoT) is one of the most hyped concepts in the computing world today. The hype surrounding cloud IoT platforms even surpasses that of IoT itself. However, both have practical applications and can be crucial for your business. In this article, we will define IoT and cloud IoT platforms without too much technical detail, and then discuss what you need to consider when choosing a cloud IoT platform.
A simple explanation of IoT is everything that is a physical entity connected to the internet. These can be sensors that measure various parameters and send data over the internet, which usually returns to a remote or “edge” server located in the same geographic area. Things in the IoT can also be directed and acted upon via the internet. Most usefully, the physical devices that make up the IoT can send measurement data while simultaneously receiving instructions.
For example, a connected “smart” soil moisture sensor can regularly report readings, and when the soil is too dry, a connected water valve opens. When the soil moisture is sufficient, the valve closes.
The moisture sensor and water valve may be connected to the same “edge computing” device or node that can communicate with the internet, or they may be connected to different nodes, as many soil moisture sensors may be used for a large farmland where only one centralized irrigation system is needed.
What is the Relationship Between IoT and the Cloud?
Of course, the “internet” is not a single terminal but a collection of interconnected networks that transmit data. For IoT, remote endpoints are typically located on a cloud server rather than on a single server in a private data center. If you are only measuring soil moisture at a few locations, deploying in the cloud is not absolutely necessary, but it can also be very useful.
Suppose the sensors are not only measuring soil moisture but also measuring soil temperature, air temperature, and air humidity. Suppose the server needs to gather data from thousands of sensors and read a forecast feed from a weather service. Running the server in the cloud allows you to import all the data into cloud storage and use it to drive machine learning predictions for optimal water flow. This model can be as complex and scalable as you want.
Moreover, running in the cloud can bring additional economic benefits. If the sensors report once an hour, then the server does not need to be activated the rest of the time. In a “serverless” cloud configuration, incoming data will activate functions to store data and then release their resources. Another function will also be activated after a delay to aggregate and process new data and adjust the irrigation flow settings as needed. It will then release its resources as well.
Feedback Loops for Local and Remote IoT
In our irrigation example, if the response time from the cloud server is an hour, the system can still function normally. Other systems may have a much lower tolerance for delays.
For example, consider an autonomous vehicle: it constantly observes the road, identifies obstacles, and measures its position. It may also continuously send data to the cloud, but it cannot rely on a remote server to adjust the throttle, brakes, or steering. These must all be done locally.
This is one of the basic lessons of control systems engineering: minimize the level of control feedback loops as much as possible. Yes, a remote manager can change destination set points or route plans, but the car itself must be responsible for all time-sensitive operations.
Basic Cloud IoT Functions
The cloud IoT platform must monitor IoT endpoints and event streams, analyze data in the edge and cloud, and support the development and deployment of applications. These are the essential functionalities required for any IoT implementation.
To achieve cloud data analytics and application development, the IoT platform also needs access to cloud storage. For industrial IoT devices and vehicles, large amounts of data can be stored, filtered, or aggregated for long-term analysis. Industrial IoT also faces challenges in network and protocol transformation. Old industrial programmable logic controllers are not suitable for Ethernet and TCP/IP.
Another challenge is how to transmit data from edge devices to the cloud platform. For indoor applications, you can typically use wired Ethernet or Wi-Fi. For outdoor applications, such as agricultural scenarios, using cellular data is common, and cellular M2M (machine-to-machine) plans can be used instead of much more expensive mobile plans.
Hosted IoT connectivity services can also help solve this issue. Some of these services mainly manage SIM cards and related data; broader IoT connectivity platforms may also involve operating systems and agents for edge devices. Note: Some mature M2M services have added “IoT” to their branding without adding any real IoT functionalities.
Considerations for IoT Platforms
You should not simply jump to an attractive-sounding IoT cloud platform; you should first determine your own needs and list some monitoring, analysis, control, and application architectures to meet them. Before using technology, clarify the design for user experience, data, and business decision-making.
Avoid designing for specific devices, device operating systems, gateways, edge platforms, networks, communication protocols, cloud platforms, or cloud brands. Instead, start with a more general design. Identify the features that are most important for your application and use that list to decide your platform choice. In other words, this will be a process.
The costs of cloud IoT can be difficult to predict and can easily be underestimated. Part of the issue is that cloud computing pricing is inherently complex. (Typically, the only way to truly understand the cost of cloud applications is to run them for a month and check the bill.) Another issue is that cloud IoT platforms often offer introductory discounts. If you rely on introductory pricing, you may be surprised when prices rise. Finally, it is also easy to overlook the costs of data storage and hard to implement a long-term strategy for discarding old, unimportant data.
Another challenging part of this process is how to assess your own capabilities. Do you have expertise in managing devices and sensors? In communication protocols and networks? In cloud application architecture, operations, and management? Are your employees able to commit to building your IoT applications, or do they have significant ongoing responsibilities? Do you need new hires? Do new hires have the right skills?
These assessments will tell you whether you need to choose a full-featured or a basic cloud IoT platform. Some vendors offer robust, nearly complete platforms that can be easily customized to meet application needs. Others provide some of the components you need but require you to do more integration and customization internally or with consultants.
For first-time cloud IoT deployments, the value of executing a proof of concept cannot be overstated. Like other projects involving software development, you need to plan for initial failures to learn from mistakes and build it correctly next time. Only after your proof of concept is successful can you begin to scale it.
Copyright Statement:This article is compiled by D1Net. Reproduction requires the source to be indicated as: D1Net. If the source is not indicated, D1Net reserves the right to pursue legal liability.
(Source: D1Net)
If you work in a certain area of enterprise IT, networking, or communications and wish to share your views, you are welcome to submit articles to D1Net.
Submission Email: [email protected]
Click the blue font
to follow
You can also search for the public account “D1net” to follow sub-accounts in various fields under D1net (cloud computing, data centers, big data, CIO, enterprise communications, enterprise application software, network communication, information security, servers, storage, AI, IoT smart cities, etc.).
