NTT’s Edge Strategy Overcomes OT Stagnation

Parm Sandhu leads the global practice for NTT called Edge and Private 5G Edge Computing. His team focuses on enabling organisations in manufacturing, energy, transportation, logistics, and healthcare. They provide processing, computing, AI, analytics, and connectivity solutions on premises. This approach differs from cloud-based services.

Sandhu explains, “The industry moved from centralised computing to distributed edge computing. We had all the server vendors delivering for on-premises and data centres. Suddenly, there was a big push back from the Googles and everyone to cloud-based computing.”

For Sandhu, Cloud offers a standardised model. Organisations can sign up and use services without understanding the infrastructure. However, he also notes that 60 to 70 per cent of workloads cannot move to the cloud. Companies need these applications to run locally. Reasons include security, data privacy, and business continuity.

Sandhu says, “If I lose the fibre, or there’s a wind storm and the fibre cables get cut, I need the applications running.”

Legacy and OT systems present barriers

Legacy systems create another barrier. Siemens and other vendors built end-to-end solutions. Infrastructure and software were sold together, creating a lack of standardisation. Vendors locked clients in with purpose-built hardware. Companies had to buy machines from specific vendors for specific applications. This created silos that IT professionals could not manage.

“The issue is standardisation,” Sandhu states. “You can’t take standard purpose hardware, containerised hardware, and just put all your applications on it.”

OT environments complicate matters further. Factory systems were built by engineers, not IT professionals. Many have been around for decades, and documentation is, at best, minimal. Systems also use protocols that IT does not understand. They rely on limited Programmable Logic Controllers (PLCs). You cannot simply upgrade PLC software without taking operations offline.

This complexity around the traditional OT world hasn’t moved much since the 70s. That creates multiple challenges that vendors have tried to address. Protocols, embedded systems, a lack of visibility and upgrade limitations are just part of the problem.

For IT security teams, the problem is even more complex. IT environments use network segmentation to prevent devices from talking to each other. OT environments use layer-based segmentation. Different layers cannot talk to each other.

What is NTT doing about this?

To provide an alternative, NTT developed private 5G networks to support this approach. These networks provide both wireless and wired connectivity and enable network-managed segmentation based on SIM-based authentication.

“You can’t do that on Wi-Fi, but you can do it on private 5G,” Sandhu says.

Traditional IT requires seeing every packet and every device. If a gateway sits between a viewer and the machines beyond it, you lose visibility. Private 5G networks solve visibility challenges by looking at the network differently. NTT and its partners integrate firewalls into wireless connections. This allows visibility into everything behind edge devices.

Next-generation firewalls include AI capabilities. Using pattern recognition, they look for anomalies such as packet transmission rates. For example, if a device normally sends 10 packets per day but suddenly sends 100 or 1,000, the firewall detects unusual activity.

Friction represents another barrier. OT professionals resist changes that disrupt operations because they see their systems as mission-critical. Any change risks the security and safety of the system.

To change that, Sandhu says that conversations should start with business outcomes. Companies should identify what they want to achieve. An example involves a legacy partner manufacturing electrical components. Their conveyor systems are critical. If a conveyor breaks, the entire factory halts. They needed to predict failures without disrupting production.

To help them, NTT deployed a private 5G overlay to place cameras on conveyor systems. “We’re not going to touch your production,” Sandhu says. “So we’ll put an overlay 5G.” The cameras capture high-speed video and still photos. AI models analyse the footage to detect cracks and fractures, allowing maintenance teams to address issues before they cause downtime.

This approach works for multiple use cases. Once cameras are installed, they can serve quality inspection, maintenance monitoring, and other purposes.

Digital Twins and modelling

One of the big changes in manufacturing and other sectors is the arrival of digital twins. They provide an opportunity to use sensor data to model a system and detect failure. Interestingly, Sandhu notes that many customers avoid the term.

He explains that traditional digital twins require expensive model building and retraining when conditions change. If a company switches from green widgets to orange widgets, vision models require complete rebuilding.

“The challenge with that was the amount of cost as you had to go through to build the models, and then you train those models,” Sandhu explains. “But then something changes, invariably changes.”

NTT has chosen a different approach with physics-based AI models instead. These are small language models that interpret physical situations through mathematical approaches. They move from deterministic to probabilistic reasoning, and zero-shot learning allows agents to interpret situations without extensive training.

For example, a forklift manufacturer. Each forklift has custom specifications based on customer requirements. Quality issues arise during production, but NTT captured 40 hours of video footage. AI agents trained on this data are able to identify quality problems and act as quality inspectors.

“You can have every quality inspector,” Sandhu says. “You train it to be that. And then, so now if something changes, it can learn.”

He goes on to explain how this approach eliminates the need for massive digital twin investments. Organisations can focus on specific agents that address particular problems. This creates value without requiring complete system digitisation.

The rising tide of vulnerabilities

Cyber attacks and vulnerabilities present a particular problem for OT. That is because many OT systems were built without security in mind. Additionally, patching OT devices, especially sensors, is not simple. They have limited memory, and it is easy to electronically kill one through a failed update.

But the biggest problem is visibility. Companies have deployed numerous sensors and controllers over the years, and often have no idea what they have or where they are. Sandhu’s team uses the Spectra Edge device to address this.

The device monitors traffic on wired and wireless networks, identifying devices by protocol and language identifiers. That ability to detect a wide range of protocols over whatever network is being used is critical for security teams. The system checks software versions against NIST standards to find outdated or insecure devices.

“If you want to start putting in sensors or a new technology,” Sandhu says. “Now you’re adding more threat vectors here to start.”

OT security focuses on controlling ingress and egress. Companies typically maintain isolation to prevent external connections. However, adding new technology requires connectivity. To prevent problems with incompatibility on the network, NTT emphasises segmentation to manage risk. Devices can only communicate with specific other devices. This limits the potential impact of security incidents.

“OT risk is usually, well, nothing else can plug into my network,” Sandhu explains. “I have complete isolation, so the risk is manageable.”

Enterprise Times: What does this mean?

Mention 5G, and most people assume public WiFi. To be fair, that’s where most of the attention has been focused by operators. With OT, 5G opened up a very different world from managing lawnmowers to manufacturing machines. But many organisations are concerned when they hear about 5G because they worry about security.

Sandhu doesn’t see that as a problem because this is a private 5G infrastructure. He believes that once organisations understand that and how to segment their network, they see the benefits of private 5G. It also provides them with an easier way to deploy connectivity across complex sites.

The post NTT’s Edge Strategy Overcomes OT Stagnation appeared first on Enterprise Times.