Engineer News Network The ultimate online news and information resource for today’s engineer
Jonathan Wilkins explains when to process data at the edge and when to send it to the cloud
Sherlock Holmes and Dr John Watson both contributed essential qualities to the duo — Holmes provided the brains and Watson kept the detective on the straight and narrow.
Just like edge and cloud computing in manufacturing facilities, the two worked together to achieve the best results.
Where does the data you collect from your factory equipment go for computation?
The cloud can be a good place to send data, especially if it requires long term storage or distribution across multiple networks.
However, as businesses are digitalising, the volume of data being sent to the cloud is increasing and this can result in slower computation and information transfer. Edge computing provides an alternative way to handle data without involving the cloud.
The benefits
In a survey of manufacturers by Make UK, 48% said they had at some time been subject to a cyber security incident.
Half of these suffered some financial loss or disruption to business as a result.
When you send data to the cloud, you create two weaknesses in security. Firstly, during transfer, data can be lost and is an easy target for cyber criminals. Secondly, the cloud is a centralised data location that is vulnerable to distributed denial of service attacks. Edge computing minimises data transfer and does not involve the cloud, so can be a more secure method of data computation.
Edge computing can also allow faster data handling because, while there is a latency when sending data to the cloud, edge devices can process data in real time.
In addition, by using edge devices to process data, you reduce the load on the cloud, meaning that cloud computation can take place more quickly. In an industry that depends on productivity, rapid data handling is crucial.
The faster data handling with edge computing means if a component develops a fault, the relevant action can be taken more quickly.
For example, if a thermocouple measuring the temperature of a pump detects an unusually high temperature, this information can be processed at the edge and the pump shut down immediately.
A replacement pump can then be obtained from a reliable supplier in less than 48 hours, so the only costs incurred are minimal downtime and the cost of the pump.
You should identify which data processing activities are on a relaxed timeframe and which require more immediate action. If you are required to respond without delay, it is best to avoid the cloud and process data at the edge.
Will the cloud become obsolete?
Cloud and edge computing each have their place in manufacturing. Edge devices eventually delete data, so any information you need to store long term should be sent to the cloud.
Also, if you need to access information remotely, it will need to be available on the cloud.
Edge computing is ideal for handling large amounts of routine data. For example, you may have a smart device collecting information on the operating conditions of your motor. Most of the data collected can be handled at the edge.
However, if the device detects unusual activity, it can automatically send this data to the cloud for storage and processing.
Just like Sherlock Holmes and Dr John Watson worked as a team to solve mystery disappearances and murders, manufacturers should use a combination of edge and cloud computing to make the best use of data and maximise their productivity.
Jonathan Wilkins is marketing director of obsolete parts supplier EU Automation.
