Edge tech a preferred computing option in manufacturing data storage

Owing to 90% of manufacturing sites not being connected to cloud or data centres, information is held in disparate, siloed systems, says telecommunications company BT digital manufacturing head Andy Rowland.

Consequently, manufacturing clients now want to benefit from Industry 4.0 and the company is finding that edge technology is often the preferred computing option to use to store client data.

Clients remain at the pilot stage of implementing Industry 4.0 and are unable to reap the benefits that it has promised, he adds.

Rowland says edge infrastructure, cybersecurity and orchestration are preventing manufacturing companies from making this step change to Industry 4.0.

Clients moving from pilot stage to l arge-scale implementation often involves their refreshing multiple parts of the network infrastructure.

Having determined how to deploy edge computing in the factory, these computing devices now need to be reliably and safely fed with data.

This often involves an operational technology and local area network (LAN) refresh exercise, as the LAN may not have been evaluated in more than ten years.

Private fifth-generation (5G) connectivity is also being considered, owing to its higher availability, better handling of connectivity in environments with lots of metal and guaranteed low latency, he notes.

However, in most cases, clients will upgrade the fixed LAN, add small gateway boxes to feed the edge compute, and upgrade other aspects, such as firewalls, says Rowlands.

“Clients state that they intend to process up to 50% of their data at the edge. They are driven by a need to reduce their connectivity costs and keep critical business and manufacturing processes running in the event of network issues.”

One of the challenges of moving to edge is setting up infrastructure that supports the implementation of the Internet of Things (IoT) solutions, as many businesses have legacy networks that cannot do this, he adds.

In addition, is the challenge of cybersecurity and safely connecting the two worlds of Industrial IoT, which is operational technology (OT) and information technology (IT). OT is about availability first and security second, and has traditionally relied on segregation for safety.

As the gaps of air between the computer and other networks are removed to leverage IoT algorithms, cyberattacks on OT will become more common.

Further, a zero-touch orchestration (ZTO) approach is the best way of overcoming the challenge of orchestration.

Prototype capabilities exist, which have been demonstrated on a full end-to-end IoT deployment, including ZTO on private networks, customer edge compute and Industry 4.0 devices.

These capabilities provide low-cost, secure and rapid onboarding, late binding capability and increased speed-to-market, states Rowland.

“Arguably, the greatest challenge is how to manage all of these edge servers. For manufacturing . . . there are more than ten-million sites globally and, traditionally, they do not have the skilled IT people who work in data centres.”

Manually installing and updating all these servers is costly, impractical and prone to mistakes, which could lead to security breaches, he points out.

Rowland highlights that the best approach to setting up infrastructure that supports the implementation of IoT solutions from legacy networks to the edge is often to develop a smart transformation programme that assesses the current mode of operation, the desired future mode of operation and the most cost-effective way of moving from “as is” to “to be”.

He points out that BT researchers believe that the ideal futureproof network infrastructure is a smart hybrid core, divided into multiple slices, that can cope with traffic being delivered over any access type, such as WiFi or 5G, based on the best available access and customer requirements.

This includes the concept of access traffic splitting, steering and switching that can make dynamic decisions about how to deliver traffic to the end-user device.

Improving the overall equipment effectiveness score is also important for manufacturing.

Some BT clients have scored lower than 50% on the overall equipment effectiveness score. This means that less than half of its equipment is either working or performing as it should or producing good-quality goods.

Improving the score using predictive and prescriptive maintenance will enable clients to produce more goods for the same outlay, vastly improve profit margins or rationalise production lines, Rowland concludes.