As industrial organisations assess and prepare to deploy IoT systems, many of the concerns about the technology, including the performance, scalability and security, are being addressed by technical innovation. However, the business model for vendors and the new value chain that is required to support virtualised technical approaches to mature industrial processes is proving harder to address.
Here, Jim Douglas, the president of Wind River, tells George Malim that industrial organisations are starting their deployments, trialling new models and new approaches in preparation for reaping the full benefits of optimised, IoT-enabled processes as the value chain surrounding new technologies matures further.
As critical infrastructure companies address the challenges posed by their legacy systems they are looking to IoT-enabled systems to give them greater flexibility while still ensuring very high availability and strong security. This is particularly apparent in the manufacturing industry in which device deployments have typically had extremely long lifecycles, as Jim Douglas, the president of Wind River, points out: “Deployments in factories have had a mentality of plugging in a device, forgetting about it and hoping for the best – with a lifecycle of 15-30 years,” he says. “The deployment cost is significant and so is the cost of maintenance but the long lifecycle and the limited functionality help to alleviate this.”
However, organisations such as these do face challenges within this relatively simple model. “Their main challenge is obsolescence and this is exacerbated because they’re not really following the innovation curve and suffer from significant vendor lock-in because changing vendors isn’t practical during a decade or longer deployment life,” says Douglas. “In addition, it’s a big step to move from the long-term deployment of tried and tested technologies onto the innovation curve which offers the potential for organisations to cost effectively move to a model that doesn’t require devices to be in place for decades to be cost effective. For example, many industrial organisations would welcome the greater interoperability offered by innovative devices as well as the means of driving down deployment and maintenance costs these can offer.”
This, though, is a big step for manufacturing organisations that are highly familiar with the embedded, self-contained environment of devices that are not connected to the internet. There’s an attitude here that the current systems work and organisations are happy with their performance, but Douglas thinks there’s a growing recognition that IoT-related technologies can enable them to access more innovation and achieve improved performance.
“There’s an attitude of: if it isn’t broken, don’t try and fix it,” explains Douglas. “Why rip something out and take on the risk of something new if what’s already in place is still performing as expected is an understandable point of view. However, these deployments are obviously not optimal if they’re not taking advantage of the innovation curve to enable effective integration of the system with others.”
“I think it’s understood that the promise here is great, but the consternation has been around the transition and ensuring that the transformation works,” he adds. “There is concern about whether the technologies out there are going to give the performance, the scalability and the security the business needs. Security is a very real concern and many organisations have equipment that probably doesn’t even have an IP address in deployment, so feel their system is wisely separate from the security challenges of enterprise IT and the internet. To move from that and suddenly have equipment that is internet-connected presents a clear and understandable point of concern.”
That concern has resulted in caution when it comes to approaching system upgrades. Organisations are looking to trial new approaches rather than commit to them across their entire business.
“You’re not going to see clean sweep upgrades, no one is going to accept the disruption of a complete deployment,” confirms Douglas. “That would be ideal in terms of accelerating the benefits, but it won’t happen in the real world. What you will see, though, is departments start to pick off various parts of their operations and slowly deploy new technologies and approaches.”
Into the virtualised environment
The other half of the equation is that companies need support from their suppliers. This is not just a technical challenge, it’s a value chain challenge. “Software-defined networks (SDN) in the enterprise, virtualisation and virtualised networks sound great on paper, but are wildly disruptive to the existing models of the current equipment value chain,” he adds. “Open hardware and open software platforms present a completely different way of capturing value and business models haven’t been defined yet. For traditional vendors, it’s terrifying. In talking to end users, we have forewarned them that you can’t get to the idealised view of virtualisation from here without addressing the value chain of your suppliers.”
Nevertheless the ultimate benefits are compelling and industrial automation manufacturers are focusing on fast wins. “If you look at a processing platform, the deployment cost of the system is explicit and rigid and, to achieve reliability the system is triple redundant,” explains Douglas. “If something fails, you literally have someone walking around to upgrade triple redundant layers in the topology. The aim is move to more standard compute platforms to aid maintenance and achieve greater flexibility.”
Yet the question of where to start remains a big issue. Companies aren’t going to implement everything overnight. “The initial area of interest is virtualizing the basic control layer. The desired outcome is adding more dedicated control loops and adding portability and system interoperability,” he says. “Challenges still remain in terms of how to upgrade, perform live patches and maintain reliability and eventually achieve high speeds and high performance going down to the lower levels of the topology.”
There are lessons to be learnt from transformations in other industries and Douglas singles out the telecom industry as one that is going through similar technological and business model transformation. “Telecom operators’ biggest issue was how to deploy services they could monetise, but the technical requirements of a telco and a manufacturing business are actually very similar,” he says. “Both require very high availability and reliability. They need the ability to maintain, upgrade and fix while the system is live, without compromising performance and minimizing downtown to minutes a year.”
In factories, the real-time requirements around actuation and control have to have guaranteed access to resources, such as for a critical safety function. CPU, memory and storage must be always available. “The notion of consolidating federated systems will also be attractive in manufacturing,” Douglas emphasises. “For example, in planes there are a large number of redundant systems and complimentary systems. Everything is about space, weight and power in planes so anything that can reduce replication without compromising safety has a huge benefit. If you can take 1,000ft of cable out of a plane, that has an enormous weight reduction impact that will have a direct effect on the cost curve.”
The capability of IoT to revolutionise the factory floor through the effective use of device-collected data is often trumpeted as a new development, but this simply isn’t the case. The data’s always been there but hasn’t been exploited effectively to date. “One of the misconceptions is that the fascination with data is recent,” agrees Douglas. “The reality is that a factory floor has always generated a ton of data, but the issue has been how to collect it and then what to do with it. A factory floor could be throwing out a thousand terabits of data per month so the challenges are how to efficiently collect the data, how to manage it, what to do with it, and how to respond rapidly to insights within that slew of data.”
“Early IoT implementations were focused on driving telemetry data from the edge to the cloud. However, many IoT use cases in critical infrastructure require feedback and decision making in real-time. Therefore, the latency associated with moving critical data sets to the cloud won’t be acceptable. This is going to lead to a lot more processing power moving closer to the edge to enable critical decisions to be made in real-time,” adds Douglas. “Decisions that need to be made in real time will leverage compute power being closer to the edge. In parallel, larger data sets can be batched and sent to the cloud in the most economic fashion for additional processing and analytics. Evolving machine and deep learning algorithms can be run on the cloud side to ultimately discover ways to improve operational aspects of deployed systems. In the long term this will be the biggest gain from IoT. If organisations can use data to better understand how to optimize system deployments, they can send control data back down to the factory floor to tune performance and system behaviour. That’s the big promise of IoT.”
Industrial IoT
As concepts of the Industrial IoT (IIoT) and Industry 4.0 gather momentum and attention, there’s a new focus on the economic benefits that industry will harness from IoT technologies, which Douglas welcomes. “We think it’s great that embedded technology is attracting attention again,” he says. “The reality is that the majority of the economic gains projected for IoT are in the industrial complex. A 1% saving on the gallons of gas consumed for airlines amounts to trillions of dollars and that’s going to far outstrip the economic value of consumer applications.”
“Wind River’s DNA is around building safe and secure, reliable systems and we have a huge footprint of customers who have been able to deploy and operate these systems for years,” he adds. “Historically, to mitigate safety and security concerns, many of these systems never touched an enterprise network or the internet. We’re in the optimum position to help customers liberate the data from these systems in a secure and cost effective manner. As part of Intel, we’re in a great position to drive complete software and hardware solution capabilities that will accelerate the adoption of IoT in critical infrastructure domains. By providing leading solutions in the cloud, the network, and at the edge we can address the entire IoT virtuous cycle.”
One critical piece of the solution includes high performance compute nodes (“fog computing”) that sit close to the edge and provide the environment for real time analytics. Titanium Control is a software virtualisation platform that enables critical infrastructure companies to cost-effectively evolve aging legacy control systems not previously designed to support the connected nature of IoT. The platform empowers organisations with the next generation of on-premise analytics so they can optimise industrial processes. Titanium Control delivers the high performance, high availability, flexibility, and low latency needed to reduce capital and operating expenses, as well as to minimise unscheduled downtime for industrial applications and control services at any scale. Unlike enterprise IT virtualisation platforms, it provides high reliability for applications and services deployed at the network edge, for example in fog deployments.
Unplanned downtime is, of course, the perpetual nemesis of productivity in all manufacturing environments so Wind River has been focusing on addressing this with its offerings. “It comes back to the attributes of platforms like Titanium Control that enable organisations to triage and operate the system live so you have the capability to perform live patching if you need a new upgrade for a device,” Douglas explains. “A lot of this is commonplace in enterprise IT, but is scary for the industrial complex. Having a system that is fault tolerant and fails over if something goes wrong is vital so being able to prove to them that you have redundancy and fault tolerance built in is one of the main proof points needed,” he adds. “Companies haven’t known technology like this exists so there is some need to inform and educate them as to the capabilities and resilience that can be achieved. With the emergence of IIoT, companies are looking to deploy nextgeneration open and secure control systems; Titanium Control addresses this need, and is in active trials with customers in industries ranging from manufacturing to energy to healthcare.”
It’s clear that we are now entering a period of sustained disruption as technologies such as IT and network virtualisation come to market at the same time as IoT offerings and alternative means of connection. These technologies are starting to find their way onto the factory floor, creating substantial management challenges and uncertainties.
“The technology is here, the economies of the technology are here but the value chain isn’t ready yet,” Douglas warns. “The approaches we are talking about are potentially very disruptive to how equipment suppliers capture value today. Breaking their current monetisation engine without a well thought out business model transition is suicide. This means they will not push initiatives that potentially render their traditional business models obsolete without knowing what their new business model is going to be.”
“I think the value chain will be the bigger challenge than the technology,” he adds. “We’ve been talking to large customers about this and they’re excited about it and looking to accelerate the speed of adoption. We’re excited about this too because we are incredibly well positioned to help companies reduce the risk of transformation so they can begin to take advantage of the incredible economic benefits of IoT. We have what is needed to make IoT deliver on its promises across the industrial complex.”
