At the edge of your manufacturing facility

edge

Vertiv’s edge infrastructure solutions enable compute decisions to be made on a site, quickly and without latency.

Manufacturers’ Monthly speaks with Vertiv Australia and New Zealand managing director, Robert Linsdell, on how edge infrastructure implemented in a manufacturing business can lead to smarter and more agile processes.  

As more devices, sensors and machines across the factory floor become connected with M2M and IoT technology, gaining insight from the data being generated becomes more important in attaining visibility and connectivity. But the data can’t be easily processed without the necessary infrastructure. This is where edge infrastructure comes into play.  

Edge infrastructure minimises the need for data to be processed back and forth to a company’s primary data centre, or a city-located colocation data centre, as it is done  instantly and more efficiently where the decisions need to be made – on the site. Processing at a nearby purpose-built, small data centre can optimise manufacturing equipment and accommodate technology integration, using predictive tools and power availability.  

Enter, Vertiv – a US$5.5 billion global pure play data centre company. It supports large cloud and colocation companies as well as enterprises in a range of industries – including manufacturing for automotive, food and beverage and semiconductors and electronics.   

edge
Vertiv Australia and New Zealand managing director, Robert Linsdell.

“Edge computing exists in multiple forms, but ultimately, Vertiv deals with edge in relation to where you process data for a specific need – close to where the action lies,” Vertiv Australia and New Zealand managing director, Robert Linsdell, said.  

“Basically, it’s where you need to have compute decisions made on a particular site that need to be made quickly and without latency.”  

In 2018, Vertiv initially defined four edge archetypes to include data intensive, human latency intensive, end-to-end/M2M latency intensive and life critical applications. This described the different data needs specific to each archetype. Upon gleaning a better understanding of the data demands their customers most needed, Vertiv recently released an updated report on Edge Archetypes 2.0: Deployment-Ready Edge Infrastructure Models. This further defines the infrastructure design for the different edge archetypes, including device edge, micro edge, distributed edge, and regional edge data centre.  

edge
Vertiv’s Edge Archetypes 2.0: Deployment-Ready Edge Infrastructure Models.

“As a factory becomes ever more dependent upon the technology that it’s using, it’s even more important to understand what infrastructure is needed to support it,” Linsdell said. “That’s why we built the Deployment-Ready Edge Infrastructure Models framework – to help people with the decision making at a simple level.”  

To uncover what your manufacturing business needs, you need to consider:  

  • business location (i.e. urban, remote);  
  • how much space is required;  
  • how much power is required;  
  • who will own it (i.e. single tenancy, multi-tenancy);  
  • external environment (e.g. will the infrastructure need to be rugged?);  
  • what quality of communication is available;  
  • should the infrastructure be passive or active;  
  • level of resilience and autonomy required for a given site or network of sites; 
  • how many deployments are needed; and  
  • will the edge solution provider need to operate their own IT application layers on the equipment.  
edge
Figure 1.

Tassal case study 

An example of how edge infrastructure has actioned positive change is Tassal, a successful salmon farming company in Tasmania, where Vertiv edge has enabled local automation of their feeding processes for the better. This has benefitted Tassal both cost-wise and ecologically.  

“For a number of years, they’ve implemented these automated methodologies of feeding and reduced the amount of food being wasted in the ocean outside of the pen,” Linsdell said.  

“They installed 4K cameras in the pens to watch the fish, so the marine biologists know the best time to feed, cross referenced with ocean currents, temperatures, weather and the fish’s responsiveness.”  

This meant Tassal was able to automate the precise amount of fish food that was required with the low-latency benefits of edge infrastructure. The fish farm now has a number of micro data centres for both salmon and prawns farming, which is a true working example of a distributed edge micro data centre infrastructure system.  

Distributed edge refers to decision making that occurs on multiple sites, using similar or the same processes, networked together to support management and communication between plants and site. Each site has its own micro data centre that leverages both ERP and MRP networked processes, bringing intelligence back to a central location.  

Linsdell explains: “Let’s say one factory produces 1,000 tonnes of food a day and another produces 500 tonnes – all that processing needs to be collated, so that they know what they’re doing and where it needs to be distributed to.”  

edge
Figure 2.

The edge of 2022 

Linsdell predicts that edge data centres will continue to be rolled out in multiple industries in 2022, having seen recent examples in sugar factories, energy companies and smart agriculture businesses.  

“If an enterprise has a manufacturing site located 500km away from a major city, there is likely a need or demand for having a micro data centre on-site to meet the desired level of autonomy and resilience,” he said.  

Additionally, the region is increasingly seeing more and more hybrid IT infrastructure, which melds cloud, on-premise, decision making, communications and data centre requirements.

“The design and architecture of the hybrid infrastructure is often a complex model. Consulting with an experienced systems integrator, in my opinion, is a key component to a successful outcome,” said Linsdell.   

“One of the real challenges for the expansion of these systems is actually using skilled engineers and technicians who have the knowledge and capability to build sucharchitectures,” he added. “And in our experience, this route of deploymentis one of the governing factors needed to help an enterprise to disentangle the many facets of a requirement.”  

Edge for manufacturers 

More than selling a product, Vertiv is focused on helping organisations to create standard processes and technologies for edge computing to help them better understand it and therefore increase its effectiveness.  

The company realised they needed to facilitate this education when 5G was introduced.  

“The real tipping point that occurred was when 5G came along, because that was able to provide you with high bandwidth and very low latency application,” Linsdell said. “So, over the last two-to-three years, 5G has served the edge story, enabling it to become real.”  

In striving to educate businesses on architecting the optimal edge computing infrastructure for them, Vertiv believes they have a responsibility to make people aware of what’s available to them.  

“One of the challenges that we see in the industry is people get very excited by the applications and all their shiny bells and whistles that all the software on the application layer provides,” Linsdell said. “But when it comes down to the actual IT infrastructure in the data centre, it can be an afterthought.  

However, consideration for where the data is housed should be an essential part of integrating these applications into your business. First and foremost, the data centre infrastructure needs to be resilient and reliable in order to support the applications that in turn keep the wheels turning in a manufacturing facility.  

“It’s about understanding what that IT infrastructure needs to look like to support the technology that’s being placed on it,” Linsdell said. “Don’t just get sucked into all that nice, shiny stuff – the consideration of that IT backbone is equally as important to the applications being chosen to drive the factory.”