The Edge vs. The Cloud: A Hybrid Approach for Manufacturing

Graham Immerman -
Illustration: © IoT For All

Edge and cloud computing are often misunderstood to be mutually exclusive but, while they may function in different ways, leveraging one does not preclude the use of the other. In fact, they actually complement one another quite well.

An Introduction to Edge Computing in Manufacturing

The edge computing framework is quickly finding its way into a variety of industries as Internet of Things (IoT) devices become more commonplace. One of the most promising edge computing use cases is in manufacturing, where these new technologies can potentially lead to massive productivity gains. 

While IoT is already proving to be a critical enabler on the factory floor, manufacturers are now looking to enhance the responsiveness of their production systems further. To achieve this, these companies are looking toward smart manufacturing with edge computing as its main enabler.

Smart manufacturing envisions a future where factory equipment can make autonomous decisions based on what’s happening on the factory floor. Businesses can more easily integrate all steps of the manufacturing process including design, manufacturing, supply chain, and operations. This facilitates greater flexibility and reactivity when participating in competitive markets. Enabling this vision requires a combination of related technologies such as IoT, AI/machine learning, and Edge Computing.

The key advantage of gathering analytics at the edge of the network is the ability to analyze and execute on real-time data without the bandwidth costs that come with sending that data offsite (to the cloud or the data center) for analysis. Manufacturing is time-sensitive in terms of avoiding the production of out-of-spec components, equipment downtime, worker injury, or death.  For more complex, longer-term tasks, data can be sent to the cloud and combined with other structured and unstructured forms of data. 

As a result, the use of these two separate computing frameworks is not mutually exclusive, but rather a symbiotic relationship that leverages the benefits each provides. 

Why the Edge for Manufacturing?

For manufacturers, the goal of edge computing is to process and analyze data near a machine that needs to quickly act on that data in a time-sensitive manner. It needs to make a decision right away with no delay.

In a traditional IoT platform set up, the data produced by a device in the field (for all intents and purposes, let’s call that a machine tool) that is collected via an IoT device is relayed back to a central network server (pushed to the cloud, if you will). 

In the cloud, all data is gathered and processed in a centralized location, usually in a data center. All devices that need to access this data or use applications associated with it must first connect to the cloud. Since everything is centralized, the cloud is generally quite easy to secure and control while still allowing for reliable remote access to data.

Once that data is processed (“analyzed”) in the cloud, which happens pretty dang quickly, it can be immediately accessed through an IoT Platform (such as MachineMetrics) in a number of ways, whether it be via real-time visualization, reporting, diagnostic analytics etc., to help improve your ability to make decisions based on real data. 

The problem: the situation gets more complicated when it comes down to decisions that need to be made extremely quickly. 

First, it takes time for data to travel the “distance” from the edge device back to the cloud. This slight delay might only be a matter of milliseconds, but it can be critical for certain decisions such as stopping a machine tool from breaking. 

Secondly, these machines produce a crazy amount of data (hundreds of data points every millisecond) and all that data traveling back and forth between the edge and the cloud strains that communication bandwidth. 

The solution: rather than constantly delivering every piece of this data back to the cloud, edge enabled devices can gather and process data in real-time right there, at the “edge” of the machine, allowing them to respond faster and more effectively.

Edge Use Cases in Action

Let’s now discuss practical reasons for the use of edge computing in manufacturing. There are a variety of business benefits to ensuring that all networks are properly connected to the cloud while also being able to deliver powerful computing resources at the edge.

  1. Improved equipment uptime: A failure in a subsystem, component or the impact of running a component in a degraded state, for instance, can be predicted in real-time, continually refined as more data is analyzed, and used to enhance operational use and maintenance scheduling.
  1. Reduced maintenance costs: Enhanced analysis of needed maintenance also means that more repairs can be completed on first visits by giving mechanics detailed instructions about the causes of a problem, what action is needed, and what parts are required—reducing repair cost.
  1. Lower spare parts inventory: Edge analytics models can be tailored to the requirements of an individual device or system. This might mean reading sensors directly associated with certain components and/or subsystems. Guided by an organization’s desired business value, the edge model can then define how the device or system should be optimally configured to achieve a business goal, making a spare parts inventory vastly more efficient at a minimal cost.
  1. Critical failure prevention: By acquiring, monitoring, and analyzing data regarding components, edge analytics can identify a cause before its effect materializes, enabling earlier problem detection and prevention.
  1. Condition-based monitoring: With the convergence of IT and OT, manufacturers are able to access machine data, allowing them to monitor the condition of their equipment on the shop floor even if they are using legacy equipment.
  1. New business models: Perhaps most important, edge analytics can help shape new business models to capture new opportunities. For example, it can improve just-in-time parts management systems using self-monitoring analysis that predicts which components will fail and when—triggering parts replacement notifications throughout the value chain. This enables the creation of an “as needed” maintenance schedule, reduces downtime and parts inventory, and results in a more efficient model.

So, when you are dealing with a CNC machine tool, in-cycle stoppages to the machine tool are an edge decision, while end-of-cycle ones can be a cloud decision. This is because in-cycle stoppages often require a very low, near-zero, lag time, while the end-of-cycle stoppages have a more lenient lag time. In the former scenario, the machine would have to leverage edge analytics when in-cycle to adapt and shut down the machine automatically in order to avoid potential costly downtime and maintenance. 

It’s Not Edge vs. Cloud…Right?

We know the point of Industrial IoT (IIoT) is to apply advanced analytics to vast quantities of machine data, all with the aim of reducing unplanned downtime, reducing the overall cost of machine maintenance, and leveraging machine learning capabilities. The cloud has been instrumental in making this kind of massive data acquisition, transfer, and analysis possible.

When data speed is the order of the day and connectivity needs to be solid, the edge will be the solution that manufacturers should look to. Applying AI and machine learning algorithms to alert, diagnose, and predict problems in real-time is a goal that can be more readily accomplished with proximity, speed, and a solid network, especially if that goal is to enable your team to take immediate corrective action or to apply an adaptation automatically without human intervention that avoids a costly failure.

To be clear, Edge computing will not replace cloud computing, though the two approaches can complement each other. Cloud computing is a more general-purpose platform for data collection, analytics, and historical reporting, but there are hundreds of use cases where reaction time is the key value of the IoT system, such as certain predictive maintenance events, where sending real-time data to the cloud prevents that analysis from happening quickly enough. 

Manufacturing companies need to be able to make decisions at three different levels: at the machine level, at the factory level, and at the business level. By incorporating edge computing with cloud computing capabilities, companies can maximize the potential of both approaches while minimizing their limitations.

Author
Graham Immerman - VP of Marketing, MachineMetrics

Contributors
Guest Writer
Guest Writer
Guest writers are IoT experts and enthusiasts interested in sharing their insights with the IoT industry through IoT For All. If you're interested in contributing to IoT For All, cli...
Guest writers are IoT experts and enthusiasts interested in sharing their insights with the IoT industry through IoT For All. If you're interested in contributing to IoT For All, cli...