IoT analytics gateway at the edge of the cloud
Data here, data there, data everywhere… and this is just the beginning of the data deluge as ever more smart devices get connected to the Internet.
Even this second, approximately 4,800 new "things", also called smart devices, get connected to the Internet of Things (IoT).
Many of these "things" reside on factory floors and connect to the Industrial Internet of Things (IIoT), together forming Industry 4.0.
But, no matter where it comes from – data is just data. For data to mean something, you need to extract information from it.
Extracting information from IoT data
A common strategy for extracting information from IoT data is to transport the data from a wide range of data sources such as sensors or smart devices to a cloud data center, run the data analysis application there, and return the results. This scalable, low-maintenance solution is perfect for analysis of historical or slow-moving data.
Yet, cloud computing might go completely awry where it is key to analyze time-sensitive data from smart IoT devices, sensors, or other data sources. If timing is key, you need fast, real-time analysis of time-sensitive data, also coined real-time streaming analytics.
Ask the important questions first
When you are planning your IoT data analytics solution, the three important questions to ask first are these:
- What goal you would like to achieve?
- What is the information you need and what data will give you that information?
- Is that data time-sensitive? Do you need to be fast to analyze it, since otherwise its value is gone? In other words, is it a real-time streaming analytics problem?
Time-sensitive data and the cloudIf your answer is “yes, my data is time-sensitive”, it’s time to ask yourself these important questions:
- Do I want to act in time?
- Does my data need to go to the cloud? All of it?
- Do I want to save on network bandwidth, storage space, and cloud computing costs?
Act in time, process close to the IoT data source
What takes less time: processing data close to source, or sending data from the data source into the cloud and processing it in a data center? The answer is easy: Data processing close to the IoT data source results in a far lower latency than transporting the data to the cloud, processing it there, and returning the actionable result.
When it is of utmost importance to process data with minimal delay, an analytics edge gateway that accepts (multiple) data sources and processes streaming data in real-time is the first choice. We designed FlexaGate FG400 IoT analytics gateway exactly for this purpose.
You might be wondering: what would I such use an IoT analytics gateway for? Possibilities are endless. Use it to analyze sensor data on the factory floor, to help cut the cost of machine failures, or to help reduce the risk for factory accidents and on-the-job-injuries, etc.
Processing data close to the source is a no-brainer for solutions that only work if data is processed in near real-time. Many industrial applications, such as for instance detecting faults during the production process, depend on image and sensor data processing close to the source. If the data couldn’t be processed close to the source, it simply wouldn’t be turned into action quickly enough.
All data to the cloud?
It may sound odd, but here's the kicker: does your IoT data need to go to the Internet? All of it? And can it?
In some cases connectivity to the internet is unreliable or limited and the only way to extract information from data is to process it with an IoT analytics edge gateway close to the data source. The filtered data which contains the relevant information can be transferred to the cloud when convenient, for storage or further analysis.
But even if the connection is reliable and always available, it might be worth to ask the question if it's necessary to transport all data to the cloud. Is it relevant to store all data, or is it only relevant to store information about trends, outliers, events occurrences, etc? Which data is valuable for later analysis? Even though it may mean that most of your IoT data never reaches the internet - as long as you derive the information that you need when you need it, you've achieved your goal of acting on your data (rather than fill the cloud).
Save on cloud costs
Analyzing IoT data close to the source and uploading the results of data analysis or data filtering ensures that cloud communication bandwidth, cloud computing, and cloud storage space is only used for meaningful data. Investing in an IoT analytics gateway once saves on the returning costs of cloud bandwidth, storage space, and computing.
The why of an IoT analytics gateway
If you answered "yes, I need to act in time", "no, I don't need all IoT data in the cloud", or "yes, I'd like to save on cloud costs", then an IoT analytics gateway is the way to go. It allows you to act timely on data, store and analyze only relevant data, and save on cloud costs.