As connected devices continue to churn out massive amounts of data for analysis, edge computing is ready to help take the strain, writes Robert Smith.
The Internet of Things (IoT) and Big Data are sweeping through the engineering and manufacturing industries, paving the way towards total connectivity.
And while the fourth industrial revolution is still a long way off for some businesses, many others have already begun laying the foundations for more sophisticated production.
Enablers driving industry towards smarter factories include improved information access from sensors and products as they are moving along the production line and throughout the supply chain.
Smart data analysis and direct control and interoperability of machines, alongside autonomous decisions being made by appliances, will also help enable the development of smart factories.
This is according to Paul Taylor, head of industrial products (UK) at TÜV SÜD Product Service, who says: “The IoT connects people and machines, enabling bidirectional flow of information and real-time decisions. Its diffusion is now accelerating with the reduction in size and price of the sensors, and with the need for the exchange of large amounts of data.”
Research organisation Statista predicts that by 2020 there will around 31 billion IoT devices worldwide, with this number swelling to 75 billion by 2025.
The IoT connects people and machines, enabling bidirectional flow of information and real-time decisions. Its diffusion is now accelerating with the reduction in size and price of the sensors, and with the need for the exchange of large amounts of data
Paul Taylor, head of industrial products (UK), TÜV SÜD Product Service
Harnessing the data produced by these devices and turning it into actionable insight is one of the major challenges industry faces.
Currently, the bulk of data processing for IoT devices happens in the Cloud. A report by market intelligence platform CB Insights says that within the next 12 months, 1.5 GB worth of data could be generated per person each day.
However, CB Insights argues that the Cloud is “no longer sufficient to instantaneously process and analyse the troves of data generated — or soon to be generated — by IoT devices, connected cars, and other digital platforms”.
Keeping close
This is where ‘edge computing’ comes in. Edge computing is designed to enable connected devices to process data closer to where it is created — at the ‘edge.’.
CB Insights says this can be either within the device itself (such as motors, pumps and other sensors) or close to the device, providing an alternative to sending data to a centralised Cloud for processing.
Localising data processing and storage via edge computing puts less of a strain on computing networks.
“When less data is sent to the Cloud, the likelihood of latency – the delay in data processing that results from the interaction between the Cloud and IoT devices – decreases,” the report says.
Picking up the pace
The edge computing industry is still in its infancy, but it is predicted the market will grow to around $6.7 billion (£5.1bn) by 2022.
Stratus Technologies, which provides technology designed to prevent unplanned downtime, is one company already providing edge computing services. This includes its ztC Edge, which is designed as a fully virtualised and self-protecting computing platform for industrial edge environments.
The company’s chief executive officer, Dave Laurello [pictured left], says one of the most exciting things about edge computing is the current level of adoption in all businesses – especially within industrial automation.
Despite this, however, there are concerns, which Laurello says is typical of any new infrastructure. For example, he says businesses are worried edge computing might not work with their systems, will be difficult to install or manage, and will cost a lot of money.
“As with any new technological development there is hesitancy to invest, and this is reflected in the industrial sector which is traditionally cautious,” Laurello explains.
“Having said that, what we are seeing is that there is a growing appreciation within industry of how an edge strategy will increase efficiency and agility.”
Laurello expects a huge uptake in edge computing as industrial organisations seek to gain increasing insight from the data becoming available at the application layer.
Hardware is one of the chief reasons edge computing is now a reality, with the latest offering being ruggedised to operate comfortably beyond the controlled conditions of the server room, and simple enough to be set up by people without extensive IT experience or qualifications.
“This means it can be put to work at the machine or line level and networked to operate with other devices, and/or into the server room, and/or into the business IT layer and up to the Cloud,” Laurello says.
For engineering firms, speed is expected to provide the clearest advantage of computing at the edge, with the Stratus CEO adding that edge computing offers a really localised place to manage data from the growing inputs available at the application level, as well as the processing capability to turn that data into information to make decisions in real-time.
“We expect the advantages of edge computing will, for most use cases, make it an intrinsic part of a computing system where it combines the benefits of the edge, ‘fog’, data centre and Cloud.”
However, while it has advantages over Cloud computing or more centralised solutions, it shouldn’t be considered independent of these approaches, since it is often complementary to them.
What we are seeing is that there is a growing appreciation within industry of how an edge strategy will increase efficiency and agility
Dave Laurello, CEO, Stratus Technologies
Furthermore, edge computing should not be considered as a replacement for Cloud Computing – rather the two technologies should work in tandem for deeper data analysis.
Laurello says: “With the increased adoption of edge computing we are estimating that 40%-60% of data will be analysed and stored locally. After the initial analysis is made, some data will be sent to the Cloud for post-screening or deeper, business-level analysis and insight.”
According to Laurello, having this level of control about where data is processed, and which data is processed locally, could offer industrial leaders the opportunity to keep business-critical data on site, rather than transmitting it to and from the Cloud, offering an intrinsic security benefit.
Security concerns?
But edge computing is by no means a silver bullet and will not eliminate issues such as cyber-attacks and data breaches.
“Cyber-attacks are a common worry in today’s IT climate. Threats like data breaches and decreased system performance can not only damage a company’s bottom line but their reputation, and in some cases, the safety of employees or the public,” Laurello warns.
However, he says edge computing can help reduce data exposure as it keeps data that might otherwise need to be sent off-site, or to the Cloud, on-site.
“All industrial facilities need to take a proactive approach to understanding and mitigating the security vulnerabilities of their processes, combining digital, physical and educational best practices to reduce the risk to the organisation,” Laurello says.
Taylor explains that as more devices are connected, and therefore more data is transferred and analysed, the risk landscape will shift, becoming more intricate and potentially more hazardous.
“Based on its complexity, a dynamically reconfigurable system requires a regulatory framework that delivers an integrated approach to safety assessment.”
To combat safety concerns, TÜV SÜD is developing certification scheme for smart factory components, addressing interface standards, interoperability, functional safety and IT security.
“The safety and security methodology represents the final crucial requirement necessary to realise the vision of tomorrow’s smart factory,” Taylor says.
