Digitization, rather like life, is difficult. Whether we describe digitization with the already over-used term digital transformation or not, the push to migrate our old and existing (third industrial revolution) business systems upwards into Industry 4.0 with cloud computing, mobile devices, data analytics and Artificial Intelligence (AI) all coalescing to create the Internet of Things (IoT)… is difficult.
Bernd Gross is senior VP of IoT & cloud at German technologyhaus Software AG. Given his company’s previously already vocal track record in IoT technologies -- and the fact that Gross heads up Software AG’s Cumulocity division (a firm Software AG entered into partnership with in March 2017 and then subsequently acquired in March 2018), he may have a valid position to offer an opinion on how to get make IoT advancements work in any type of business scenario.
“There is a tsunami of new devices and IoT ‘things’ coming to the fore because the cost of connecting those devices is so much lower today. But over 50% of IoT projects fail… and one of the major reasons for this is that projects try to run before they can walk in terms of assessing connectivity factors and the total scope of the technology engineering involved,” said Gross.
Gross suggests that IoT success is only possible (or at least most likely) through a software platform driven approach (he would say that, his firm sells an IoT platform), so what does this actually mean?
A service wrapper for IoT
Cumulocity enables Software AG customers to build what could be called a ‘service wrapper’ around each of their IoT enabled products -- and this means that they themselves can deliver new services to their own customers. Those 'wrapped' services could be anything from a printer that knows it is running out of ink to a piece of IT hardware that knows it isn't being used properly... all the way through to a bus that knows its engine is producing the wrong kind of emissions. That information emanating from those IoT device services can then be fed into business dashboards. Those dashboards could be (and Gross uses an actual real world example) a graphical screen showing the performance status of a collection of industrial compressors. In the background, Cumulocity allows companies to apply a layer of business logic to the Key Performance Indicators (KPIs) being tracked by the dashboard – and all that functionality is available to software developers through open programming Application Programming Interfaces (APIs).
Software AG says that customers and developers can use these interfaces to extend the core functionality offered, or even write entirely new applications. This, insists Gross, enables the business to focus on its competitive edge (and look at product differentiation and new innovations) instead of dealing with complex basic infrastructure such as scalability, security or multi-tenancy.
“An industrial wind turbine can generate up to one Gigabit of data per day, so when we started to build Cumulocity we knew that we needed to build not just an IoT cloud platform, but also something that could run at the ‘computing edge’ as the amount of data on the device is so high,” said Gross, explaining the trend so prevalent in the industry to push a greater degree of intelligence into IoT devices themselves, so that not all processing has to happen back in the datacenter.
Software AG’s VP of industry marketing Theo Hildyard backed up theories put forward by his colleague Gross, which offer three major procedural plays, steps, stages or phases (call them what you will) that could help firms start to plug in the sensors and tracking devices that the IoT offers into their business plans.
Three procedural IoT plays
IoT Stage 1 – The best place to start is with discrete defined projects where one single IoT application is created.
“Sensors and other instrumentation will be put in place – and some of it will may need to be retro-fitted. A proportion of the data being analyzed may be historical data, but equally, some of it may be live streamed into the application system, so a degree of device management will need to be put in place. During this stage, the business must confirm the commercial validity and operational efficacy of the project. There’s a further de-risking element here too because the stage #1 IoT application will not necessarily be fully integrated with the rest of a firm’s live operational processes,” said Software AG’s Hildyard.
IoT Stage 2 – This stage has been called the period of data-driven process integration. Hildyard explains that this happens after stage #1 has been completed and represents the point at which a firm can start to write the results of IoT analytics into the way the business operates. This is IT/OT – Information Technology Operational Technology convergence – the point at which IoT starts to be truly plumbed in to the business. As TechTarget notes here, OT systems are used to monitor events, processes and devices and make adjustments in enterprise and industrial operations.
“This stage starts to involve and impact people in the workforce, so it needs careful consideration. At this point we can start to use IoT intelligence to perform actions such as a) ordering an engineer if maintenance is needed or b) initiating an order to restock a ‘machine consumable’ such as oil, or paint or some other core commodity that the equipment itself uses up as part of its core operations,” said Hildyard.
IoT Stage 3 – In this final more advanced stage, we can start to learn and innovate as we apply machine learning and Artificial Intelligence to the business. This is all about building up a body of knowledge so that the business can start to apply IoT analytics into the organization. This is the point where firms can start to uncover new products and service opportunities and even break into new markets.
Software AG Cumulocity’s Gross also outlined three other major considerations when selecting an IoT platform:
- Ability to run rapid deployment - this is important because with so many IoT projects failing, it’s important to be able to move quickly and start an idea within days rather than months… as soon as it is stable, get it deployed.
- Open independent technology – all the functionality should be accessible through open APIs, not just 10% of the functionality… we need to be able to extend the project out to a wider operational level if it is successful.
- Secure robustness – there is a lot of focus on security and data privacy as an end-to-end consideration, so we need to bring this forward as a key concern. This is not just technical security, this is also physical security of actual devices out in the field.
Returning to Software AG's Bernd Gross, who does he think is the firm's biggest opponent?
"Our biggest competitor is any self-made system. Although bespoke custom-built systems will initially be tuned for the job in hand, they will always lack an inherent openness. This means they will always be tougher to adpat when business models change, as they always inevitably do. Custom-built IoT software eventually falls foul of its inherently inconsistent architecture that has been changed too many times. With an abstraction layer and IoT platform approach, you always have the flexibility to handle each adaption as a separate service, so the end result if more control," said Gross.
A new notion of the IoT
Have we developed a new understanding of the Internet of Things? Have we got past the initial realization that the massive streams of data emanating from the IoT is a ‘drinking from a firehose’ situation? Does the platform play act as a cure-all to allow us to stand back from the data torrent? Is the three-stage approach to careful and considered IoT project development sensible?
It’s tough to answer all these questions, apart from one perhaps. Taking things strategically slowly but doing so within the context of open technologies for interconnectivity is hard to argue against. Perhaps it’s three steps, perhaps it needs to be broken down into four, five or more. However we map it out, bringing IoT online needs to be planned and measured with every new data node being validated as carefully as a new human employee.
Firms everywhere in every business vertical now have an information security policy. These same businesses are now developing voice, chatbot, AI and machine learning polices to govern what they do with these new technologies. The company IoT policy is around the corner.
