Guest Blogs

Dissecting Customer Data processing in IoT for Enhanced Marketing

Customer Data is the most significant aspect of marketing & subsequent Customer Experience. Explore the anatomy of IoT data processing & IoT cloud data storage.

We hope you have a basic understanding of public cloud-based systems such as Amazon, Google, and Microsoft Azure.  And also having a basic understanding of data and databases on how information is stored and processed. You need to be familiar with how applications are built, deployed, and maintained.

Basics of IoT:

The very common question is, “what is an IoT Device? Well, it is any device, typically not a computer or a server that will store and process data and transmit data to some sort of a centralized storage system. We have millions of these today and perhaps 100 of them that we see in our life daily. There are phones, cameras, drones, smartwatches, automobiles, etc. like anything that can absorb and process data and retransmit it to some sort of a central device.

IoT devices are all around, and they are systemic. Ultimately this is about connecting things. And it is great that we have some device that can process information such as a smartphone or even computers within the car, but the ability for those devices to transmit information to some other type of a central computer and aggregate the data and share the data and respond to the data is where the value of the Internet of Things comes in.

So, ultimately, this is going to be about not only dealing with the devices that pick up the data and transmit the data to a central source but what happens at the central source especially in the cloud. Cloud Computing becomes probably the most viable topic and target that the devices can connect with. Therefore, IoT and Cloud Computing are tightly coupled.

IoT Data Flow:

  • The IoT specifically how it works with cloud computing, is all about data production and what we do with that data. So, here in this article, I am going to explain about devices that send things back to a cloud-based system and a cloud-based system that may intern control aspects of those devices. 
  • For example, a drone transmits information to a central computer on a cloud. The drone can now contain data that is the core of the drone, such as the ability to respond to objects such as navigation data and the object in front of it, so that it can avoid collisions. But the ability to take images and videos from that drone for example, over some sort of Farmer’s field to look at water saturation, and transmit those images to a centralized server is really what the power of IoT is.
  • So, this is about leveraging the data. We may have local storage, which occurs at the device, and certainly, with the power of nonvolatile memory today, we can store gigabytes of information on devices that can fit in the palm of your hand. But this is also about storing information in some sort of a remote System, typically exiting on cloud computing. But it can also be a proprietary on-premise system.
  • So, in some places, it has some sort of a network connection that can receive, store, and organize data. So, cloud databases tend to be the preferred technology for this because they are already connected, they’re pervasive, and they can scale up to as high as we need for them to scale up.
  • But we have to do something with that data as well as to dell with the business intelligence service which, is the ability to look at the information that is coming back from logistics. For Example, in terms of how efficient truck drivers are moving across India, based on devices that are placed in the truck which, are IoT devices that transmit data back to a centralized server

 

How Data is processed in IoT Devices:

  • As we discussed the focus of IoT is data. So, now how do we leverage IoT data from say, a factory device as an example. So, we have IoT data that comes from that device and it can be any number of things in terms of maintenance, scalability and in terms of how efficiently the arms are processing, how much heat is being spun off, etc. any number of things that are pertinent to the maintenance of a robot.
  • So, therefore, we have some sort of operational understanding as well as analytical understanding. We can look at the console of Robot and figure out the temp., the efficiency, the power consumption, and also, we can look at the data in some sort of analytical way. In other words, looking at how the temperature, operational characteristics that we would be always gathering from the robot. Looking in the context of a larger analytical database that may track the maintenance of robots. So, we can be able to look at the operational instances of that robot.
  • For example if the temp. of the robot has been too high for a couple of days and we know that it may be time for a new cooling system to be installed if we may figure out that the coding system is not necessarily going to solve the problem if the temp. in the factory has been going up over time. All these things need to be taken into consideration.  Need to be able to log data from an operational standpoint.
  • So, if we’re gathering data, we’re looking at what data led up to some outcome that occurred, led up to failure or led up to success and ultimately if we are getting an outcome i.e. failure or success, we’re looking at a number of parameters. That is the matter of looking at operational data.
  • Analytical data is all about prediction, the ability to use historical data. In this case maintenance and outcome for factory-level robots so that we can prevent operational outages from the robot because we can proactively able to figure out issues before they become issues.

 

  • So, you can leverage those predictions to make sure that the robot operates at the highest efficiency without any number of outages. If an outage does occur, then how to repair the robot in a very effective way? 
  • We are also externalizing IoT data to dashboards and that’s typically going to be for human consumption. So, these may be on a console in the device itself in the case of a robot or this may be something i.e. externalized to a cell phone or any number of ways, but they provide the meaning of the information that’s coming off of the devices. In some way, humans can consume them and react to them. So, it is very much like a business intelligence dashboard.

Data Storage in Cloud:

If you are aware of how edge computing is related to cloud computing and IoT in general? Ultimately this is balancing processing at the IoT device with the centralized server exiting in a cloud-based platform.

  • The reason we move some processes to the edge or within the IoT devices is ultimate, it may be more effective and faster either from a cost-efficient basis or just from an overall efficiency basis.
  • To keep the information stored on the device such as within a Jetplane, as to the ability to react to some sort of engine issues and so while it is good to transmit that information back to some sort of centralized server so that we can trend that information over time. 

  • It is going to require that people on board and computers onboard take note of that issue and repair it ASAP. It is about moving data and processing much closer to the source than if we sent everything to the central server. So, it’s good to use cloud computing for higher-end processing. In other words, the things that require a lot of horsepower, such as predictive analytics and
    machine learning technology

  • But it is not necessary to transmit every piece of data from the IoT device  to the cloud and therefore some of the processing can occur at the source or within the computer within the IoT device. 
  • So, the ability to let the IoT device take some operational processing is able to provide some benefits. For example, If the network goes down and if we disconnect from the central server, then that device can continue processing for a time. Even though it cannot send data to be trended at the central server, it can still react to issues it needs to react to.

Edge Computing Summary:

  • In the IoT devices, Edge computing balances the processing
  • It moves the processing closer to the Source
  • It allows an IoT device to take on any operational processing

Edge-based and Cloud-based Systems Commonalities:

  • So, an example with Edge computing and a cloud-based system is really kind of interesting. They both have a couple of things in common like they both have Operating systems typically the edge-based device and it could be an IoT version of Linux or Android OS. And the Cloud could have an operating system as well typically Linux or WindowsNT.
  • They both leverage data storage and they both use application programming interfaces to communicate inside and outside of themselves.
  • They both have applications that run on them. So, remember that an Edge-based device, even though it may be inexpensive and typically not as powerful as a server that sits on a cloud-based system, it is going to have a lot of the same characteristics. It’s going to have the ability to store data, to process information, and respond fairly quickly.
  • As the technology matures and progresses, the edge devices are becoming more powerful. You can just look at the evolution of the cell phone. So, the cloud provides the advantages of scalability. So the reason that cloud is even in this equation is that we know that cloud won’t run out of the resources, so if send 10 petabytes of data which is spinning off an IoT system, we may have a million devices out there that are gathering the information and even though they are processing some of the information at the device, in any edge computing type fashion, most of the information is going to be sent to some sorts of a back-end system.
  • We’re typically going to pick a cloud computing system to do that because of the elasticity or the scalability. In other words, it’s typically not going to run out of resources because the resources are virtually unlimited. We can provide as much storage and as much compute as we need, so that is why we do so. 

IoT & Cloud:

Why are IoT and public clouds coupled? Why are they dependent on one another?

  • Remember that public clouds ultimately have certain patterns or things they do very well. Examples would be AWS, Azure from Microsoft, and google cloud platform. There are a number of other public cloud platforms as well. There are specific IoT systems or services on each of the cloud platforms.
  • For example, on AWS, we have AWS IoT and they provide AI capabilities and also analytical capabilities that are native to that cloud. So, they know that IoT is going to be an important part of computing and they have certain services out there that are able to maximize their ability to provide value to IoT based applications.
  • On Azure, we have Azure IoT as well, artificially intelligent systems, and also big data analytics. They may be different in brand, but you will find that a lot of patterns are exactly the same.
  • In Google, we have Google cloud IoT core and AI as well as data analytics which also builds into Google platform. So, which ones you leverage really should depend on your requirements and remember that there is no hard and fast rule about your leveraging one.

Example:

 

  • We may have an IoT based application that is built in a particular drone that may be overflying a farmer’s field to determine soil hydration. And the device itself has its own sets of processing such as navigation and keeping the drone in flight as well as when to take pictures and what to take pictures of.
  • We are transmitting back to some sort of public cloud service which is aware of the IoT system, in this case, the operating system and the format of that system, and even has prebuilt applications for leveraging image processing. In this case to process hydration saturation for a farmer’s field. So, they can work and play well together.

Useful Cloud Services:

  • Artificial Intelligence
  • Business Intelligence
  • Data Storage and Processing
  • Service Integration

Artificial Intelligence

  • The device is aware that the public clouds are around, and the public clouds are aware that the device is around. And obviously, the ability to support a specific device dependent on the public cloud that you pick.
  • So, cloud services include artificial Intelligence-based systems or machine learning-based systems. Those are how we can make sense of gobs of data that are coming off of the IoT devices. 

Business and Market Intelligence

  • Business Intelligence or the ability to do things such as predictive analysis, to predict maintenance cycles or really kind of the predictive outcome of information that we see coming off an IoT device. Such as health informatics or information that is coming off of some device i.e. attached to our bodies that need to alert the physician because some sort of issue is occurring.

Data Storage and Processing

  • Data, the ability to have databases that are able to deal with petabyte plus amounts of information be able to process that data quickly and efficiently. And also be able to make sense of the data no matter whether it is structured or unstructured.

Service Integrations

  • Service integration means the ability to integrate all the services together. Remember that a core feature of linking cloud-based systems to IoT based systems is the ability to efficiently integrate those systems.
  • Obviously, systems that are out and about on very weak WIFI network in some instances, communicating with Bluetooth in other instances, an older networks that may be a part of a factory, have to have some resiliency built in them in terms of how information is going to be processed from source to target be able to deal with all sorts of issues and do so inefficient ways.

IoT Data storage in the cloud

Example scenarios: 

  • Suppose we have many sensor devices and all these sensors collect data from the various resources and these sensors forward the data on the network to a particular computer/Mobile device that demands the sensor data. And then that specific computer/Mobile device uses the DSP to store the different types of sensor data on a local database.

So, just to explain to you how this DSP internally works?

Any device like mobile or desktop or any sensor that can use the IoT device through the client SDK or Mobile Apps or website etc. Then the IoT devices will respond to the caller devices and will collect information from various sources. DSP will take care of the following things:

 

  • Security Controls
  • Scripting & Customization
  • Auto generated rest API

Security Controls: Every device must have the proper authority to access the particular data and also DSP takes care of Authentication as well.

Scripting & Customization: This means what type of data a particular user wants only that type of data should be provided and also in terms of service permission like which types of services to that user is not granted or granted permission. Accordingly, the user will be provided access to that particular device.

Auto-generated rest API: This will generate the APIs required to store the data on the client database as well as on the cloud database. Also, DSP provides a secure connection between the Rest APIs and the databases. Storing the data in local DB or cloud DB does not matter for DSP only the rest API would change.

Conclusion:

Here we have explained the very fundamentals of IoT Devices and how data is processed in IoT systems. Also, we have mentioned the tight coupling between IoT and cloud computing which would give you clarity to understand the high level overview of IoT systems and data flow. 

Along with that we have also tried to touch base on the use of customer data for Iot and Retail Marketing and how IoT cloud data storage is an important factor in facilitating the relative ease for the entire simulation. 

Check Out The New Martech Cube Podcast. For more such updates follow us on Google News Martech News

James Warner
Senior Business Intelligence Analyst at NexSoftSys
James is a Business Intelligence Analyst as well as Experienced programming and software developer with Excellent knowledge on Hadoop/Big data analysis, Data Warehousing/Data Staging/ETL tool, design and development, testing and deployment of software systems from the development stage to production stage with giving emphasis on Object-oriented paradigm.

Previous ArticleNext Article