When launching several smart, connected products based on the same IoT architecture, is it better to manage all of them on a single IoT platform or to make use of several platforms?
The answer, which I’m sure will astonish you, is… it depends. A deciding factor is the degree of similarity between the smart, connected products. Here are three different scenarios to consider:
1. Managing similar products with a single platform
When managing similar products, using a single IoT platform is generally the most efficient approach. Let’s take smart wind turbines, which generate both electricity and lots of data, as a case in point.
Because wind speed and frequency varies from place to place, wind turbine makers offer a variety of products that differ in their physical parameters and the volumes of energy that they can generate. Yet, despite these variations, they will all gather, store and processing sensor data and send commands to actuators in a similar manner, and the cloud receives sensor data in the same way regardless of the turbine type.
Also, since most of the sensors collect data about air pressure, temperature, vibration, tilt and other shared turbine characteristics, they are all quite similar, and the data warehouse can use the same tables for storing data about different turbines.
What this means is that a single application can run on several types of turbines, although portions of its logic will be specific to each product. And since the values that guide the choice of the commands for the different types of turbines can also be stored in the warehouse, different control apps can make use of the same tables.
All this is helpful for controlling the performance of an IoT system and detecting any operational and security breaches. Data analytics can be performed both separately and jointly to either evaluate the performance of the same type of turbine or compare the performance of different turbine types.
Users may be either be equipped with several control apps for each turbine type, although they may share a common user interface, or with a single control app for all of the turbines.
An exception are machine learning applications, which in most cases should be conducted independently for each type of turbine. But some overlap and data sharing is still possible, for example by tapping data from older products when a new product that has yet to generate any data is launched.
2. Managing dissimilar products with a single platform
Now let’s consider when it’s possible and beneficial to run different products on a single IoT platform, using GE turbomachinery systems for the oil and gas industry as an example.
The company’s product portfolio includes drivers (aero-derivative gas turbines, heavy-duty gas turbines and synchronous and induction electric motors), compressors (centrifugal, axial, direct-drive high speed, integrated, subsea and turbo-expander) and turn-key solutions (industrial modules and waste heat recovery units). These products are entirely different from one another and are used at consecutive stages of the oil and gas mining process.
But while the information gathered varies from product to product, sensor data from all of the products can be stored in a single data lake. There is no risk that the data from different products will be confused, although separate gateways should be used to filter the temperature, humidity and various other readings that come from the various types of equipment.
Also, the data warehouse should store the data in separate but interrelated tables, as each product has its own specific sets of parameters. This allows the sundry products to “cooperate” and share some data.
Different control applications are also needed, as each product performs different operations, but a single, well-tuned IoT system can be used to automatically control the entire oil and gas mining operation.
The initial data analytics should be conducted separately for each type of machine, but afterwards, the results can be merged to search for common tendencies and patterns.
Another possibility is to manage all the products through a single user application that provides separate views for each piece of equipment along with an integrated view of the entire mining operation.
3. Managing dissimilar products with different platforms
Finally, let’s consider the case where different products work better when connected to different platforms. For example, an IoT-friendly energy provider with both wind and water turbines would likely utilize a separate platform for each type of turbine.
The reason is that, unlike the previous example, in this situation the different types of turbines are not linked, and connecting them to the same platform would result in few benefits but much greater complexity, since different gateways, data warehouse tables and control apps are required for each type of equipment.
Pluses and minuses
While there can be many advantages to running an assortment of smart, connected products on a single IoT platform, there can also be significant drawbacks. The benefits of using a single platform include:
· Software reuse, since the same cloud software may serve for various products.
· Ease of learning, since it’s faster and easier to master new products when they are connected to a familiar platform.
· Ease of operations, since managing, monitoring and controlling devices through a single IoT platform avoids the additional connectivity and compatibility issues arise when working with different platforms.
But common disadvantages are:
· Greater complexity, since different connected products may require different algorithms, data warehouse tables, control apps and more. When different devices don’t need to work together in a coordinated fashion, the wise move may be to run them on separate platforms.
· Incompatible data, since different products generate different data sets. Given the amount of data that needs to be stored, processed and analyzed as the number of smart devices grows, managing it all through a single platform can become a real stumbling block.
The bottom line
A single IoT platform makes it easier to reuse of software, share experience and integrate the output from different devices. On the other hand, disparate products require different algorithms and controls, making it simpler to manage them from separate platforms.
Since the issues that arise are often more complicated than they first appear, weighing the pros and cons of using a common platform and settling on the best approach is one of the most important pre-conditions to effective IoT adoption.