Mission-critical Edge Processing – A key requirement for IOT data

Mission-critical Edge: Internet of things is growing at phenomenal phase. The nucleus of IOT is the data, and as IOT is expanding, so is the data. With the growth in IOT, a sensor is attached to any, and all devices. These sensors keep churning out the data every second; what we are talking about is trillions of Gigabytes of data generated by IOT devices. Cisco predicts that by 2020, IOT will generate 600 trillion GB of data. 

These massive volumes of data will drive data management challenges for IOT. In the current architecture, the data collected by these sensors are all moved to the cloud. For example, in the case of traffic management, the data from different sensors across the city are being routed to multiple clouds sometimes to partner cloud. The data in the cloud are analyzed, and some anticipated events are driven based on the processing. The real potential of IOT is when these data can be used to create real-time actions. For instance, in the case of traffic accidents, the data collected by street sensors need to be processed in real-time. After that, only it can give live feedback in re-routing the traffic in avoiding congestions. In these scenarios does it mandate to loop the data to the cloud sometimes to partner cloud and back. The real value will be to process the data, analyze, and drive action at the Edge for quicker response time. In situations where the data changes spontaneously warrant an instinctual response. 

So the data management in mission-critical IOT application becomes crucial. Moreover, Edge is how data management can be made efficient in IOT mission-critical or not. Let’s take an example of non-mission critical use, an IOT smart home device like a thermostat does not need to send the data to the cloud every time. Edge has to make intelligent, the intuitive decision on data routing to the cloud. Edge has to determine whether collected data needs a real-time response. Hence data will have to be crunched immediately or is more of a storage and pattern collection data, thereby sending to the cloud. Also, if it is routing the data to the cloud where and which database does it go. Since only the necessary data travels outside the edge network, it is less likely to get into the hands of the wrong person.

So it is clear that for the best approach in edge-cloud architecture IOT data management tasks should be handled by Edge.

When talking about data management, we need to understand the type of data that is produced at the Edge. First this the data which requires a real-time action so let’s call these real-time data, these needs to crunched instantly and action-driven immediately. The second type of information is the data which needs to be stored in the Edge for a short time, and then the data can be sent in batch to cloud for later analysis.

For Edge to support data management, it should support storage which means it should support database. These databases have to be specifically designed for edge scenarios since these have to be efficient and will be built in a constrained architecture. Building data governance at the Edge helps in deciding which data needs to be stored and where it needs to stored and thus saving costs on storage. Building AI on the Edge helps to drive intelligent decision making at the Edge in real-time. However, at the same time, to better utilize the constrained environment, it is better to have the analytical engine on the cloud.

Mission-critical Edge: We are reaching a stage in the IOT and Edge revolution. In this revolution, data from the Edge to the cloud is refined and enriched. A robust working architecture of edge-cloud helps in improving the efficiency of the whole IOT platform.


  • Mobodexter, Inc based in Redmond, WA builds IoT Edge solutions for enterprises applications on Kubernetes & Dockers.
  • Register now at developers.paasmer.co and try a free Edge trial license of Edge Intelligence software that runs on Raspberry Pi.
  • Follow our all our weekly IoT Blogs:
  • Our IoT Newsletter reaches 3000+ subscribers– Subscribe Now.
  • Become our affiliate partner today.
%d bloggers like this: