Solving Emerging IIoT Environment Challenges with Intelligent Edge 

27 September 2021 | 4 minutes

By Przemek Tomczak

Many systems and technologies can comprise an organization’s value chain — from statistical process control, quality control, and asset monitoring and management to fault detection and classification, digital twins and simulations.  Perhaps unsurprisingly, demand for Intelligent Edge systems — which ingest distributed information and enable localized decision making — has compounded; in turn, the strategy and orchestration of these technologies has become even more critical. 

A recent survey of 700 enterprises by Forbes Insights found that edge computing has continued to rapidly expand as part of corporate computing, with 84% of respondents saying their IoT networks have grown over the past three years. The investment is paying off — more than three-fourths of IoT leaders say IoT leads to increased revenues or profitability and those who have adopted edge-computing and IoT implementations are seven times more likely to see high rates of growth. 

But there are many challenges; before businesses can realize these benefits and embark on their industrial digital transformation, they must consider:  

  • Data Capture: The volume and velocity of data are growing astronomically. Today, sensors measure against chemical changes, image, audio, speed, vibration, humidity, pressure, voltage, amplitude, location — you name it. What’s more, these measurements are typically high frequency (1 Hz to 100 KHz per sensor). 
  • Data Access: Availability is key. To support these applications and use cases, sensor and related contextual data must be ingested, processed, and analyzed in the right place, at the right time and provided to the right people. 
  • Analytical Efficiency & Impact: In many cases, data processing and analytics lean on information taken from the device, asset, factory, vehicle, or remote site, and leverage cloud or central services to augment what is happening in the field.  Although cloud and central services are important, we’re finding that they cannot deliver the insight and action required fast enough. When microseconds and milliseconds matter, moving data to the cloud and a recommended action back to a remote site quickly may not be possible.  
  • Quality of Insights: Common processing and analytics can help inform whether a part is bad, check against pre-defined control limits (and update limits), and predict when an asset will fail. But in order to do so effectively, systems must be optimized to analyze data –  be it sensor, process or asset performance — both in motion and at rest, and produce insights in real-time at the moment it matters. Joining asset and sensor data to MES, ERP, and Asset Management data sources can also help detect faults and classify them for corrective and preventative action. 
  • Data Management Cost: Simply put, most IoT platforms and analytics solutions were not designed for IoT use cases. They have been limited to store (or log) raw data, support limited enrichment and analytics, and perpetuate a cycle where valuable data is pushed to central facilities away from where the action is. This approach to management and storage is not only out of date, it’s costly. 
  • Infrastructure & Energy Consumption: Greater attention is being placed on workload efficiency, especially in resource-constrained environments (CPU, RAM, storage, power, and cooling). While on-premises data centers can be useful, they can’t be scaled down to run in a vehicle or device or a factory. They just require too much infrastructure and energy. For example, it’s not practical to deploy racks of servers in a remote location or a high-powered server in a vehicle to power analytics and underlying database technologies. 

How to Unlock More Value from the Edge 

At times, decision making can tolerate a reasonable delay — but this is not true of driving, for example, or on the factory floor. These decisions can’t afford to wait for external systems or centralized services to provide the data needed to determine next steps, and with a competitive edge system they don’t have to. As the number of connections (and data) between systems continues to grow, consider these questions to unlock more value in your operation: 

  • What outcome can be improved based on information and insights locally? 
  • What is the current level of connectivity to various sources of data, including both legacy and streaming systems? 
  • Have we defined and determined the analytics and algorithms for delivering the desired insights? 
  • Do we possess a light-weight engine that can ingest, process, analyze the data to deliver the insights quickly enough for action? 
  • Are these insights delivering expected results, and are we updating and re-training models based on new information?  

Remember, more than ever, time and insight are of the essence. Because while the decision itself might be perishable, the impact of it is not. 


This post was originally accessible to registrants of Industry of Things World USA. To learn more or attend future events, please visit:  

Demo kdb, the fastest time-series data analytics engine in the cloud

    For information on how we collect and use your data, please see our privacy notice. By clicking “Download Now” you understand and accept the terms of the License Agreement and the Acceptable Use Policy.