The ‘Ouch’ in IIoT

Pratima H

INDIA: We saw how prolific is the industrial playground that IoT (Internet of Things) is staring at in the first part of this feature.

Prolific and yet barely scratching the surface - to be precise.

Picture this. Analyst firm IHS, which has been covering this domain rigorously, predicts the shipment of internet connectable device to industrial sector will grow at 25.7 per cent from 2013 to 2025. Here, industrial automation seems to be leading the growth, and outperforming average total sector growth during the forecast period, with the shipment CAGR of 35.3 per cent.

So, if the popularity of Industry 4.0, and consequentially, the importance of IoT in the industrial sector is getting emphasized more and more, what’s the trouble?

In fact, Arun Suresh, Analyst – Data Centers, Cloud & IT Infrastructure for IHS Technology points out that this integration of networks and software with hardware and equipment not only bridges the gap between controllers and end node devices but also serves as a platform to digitalize some applications like real time remote monitoring, data analysis and preventive maintenance, which ultimately optimize the process and improve the productivity of manufacturing operation.

That's not too futuristic any more.

Swing to a 2015 Forrester survey, and we see that 23 per cent enterprise telecom decision-makers nodded to having implemented some form of IoT and 2016 could see that number going to top one-third of global enterprises. The number of enterprises planning to implement will move up from 29 per cent to more than one-third, as Frank E Gillett and other analysts observe.

That would hint leaving a lagging one-third that will come along more slowly. What is keeping this one-third on the fence, shy from playing with sensor technologies and finding new ways of troubleshooting, predicting failures, remote upgradation and increased automation?

More so when industry players like Sachin Kurlekar, ‎Associate Vice President at Persistent Systems, explain that these outcomes, in turn, contribute to reduced downtime, increase operation efficiency, reduced costs, maximizing machine performance?

If IIoT has enough use cases and applications to justify its need in industries shouldn’t everyone be jumping on to this de rigueur then?

Not yet. Not until……

Turbulence 

IIoT cannot chase the whopping trajectory that awaits it till a well-rounded orbit is in place. In other words – anything and everything that can support or impede IIoT – read backend networking gear, sensor technologies, ecosystem elements, standards, enough interoperability for IIoT to actually function out there between heterogeneous factory fixtures, the right location, context and relevance of data being squeezed real-time, and a lot more.

The tremendous potential of IIoT-backed applications in industrial scenarios notwithstanding, presently, the question is less about justification of use cases but more about whether a right solution exists. Ask questions like - What is the cost of deploying/maintaining it and can the solution be supported for many years, Kurlekar pauses to ponder when it comes to IIoT’s real assessment.

Enough accoutrements on the side, a reasonable degree of connect with legacy lay-outs and the right protocols, should constitute the backbone for such future-forward scenarios.

Vikas Chadha, MD, Honeywell Automation India Ltd. candidly confronts the challenges that any new technology disruption has to evolve through. "The IIoT in many ways is pretty undiscovered at this stage and is waiting to be explored and mapped out completely and we at Honeywell, have the technologies to be compatible with backward to old components. It is just a matter of time that customers will play a huge role in forcing vendors to supply the protocols of this new world.”

As to Kurlekar’s assessment, each IIoT solution provider uses or develops a proprietary solution which means no customer is sure if the proprietary solution can ever talk to other solutions, how it compares with other solutions and the customer gets locked in with that solution provider.

Reliability and longevity of connectivity solutions, stays a concern. Suppose some IIoT solution deployed uses 2G connectivity for data connection & 2G is phased out or the service provider stops supporting it over the next two years, in such a case, what happens to the solution or what if the telecom operator goes out of service, he questions?

Another question mark that begs attention here is the stark divide between Operational Technology (OT) and Information Technology (IT) in heavy-industries that has often hand-cuffed new disruptions where the OT-IT marriage is an imperative to say the least.

Chadha addresses the divide by suggesting small solutions instead of full-blown ones. “It would call for talent development and other capabilities to be taken care of. Different customers need different solutions so the challenge can be tackled gradually.”

Right now IIoT is very fragmented with a vertical siloed solution approach: Kurlekar

Most industrial customers are used to deploying technologies/solutions which last for decades. The current IIoT solution providers are focussed on deploying solutions which span two to three years, says Kurlekar. “Some of the companies formed in the IIoT space are even start-ups raising concerns about how long they will remain in business. Similarly, technologies are changing rapidly so predicting how current technologies will evolve or remain relevant for a decade once the system is deployed and how they will get support in the future is also a concern.

While all these doubts surround IIoT, the data part , supposedly the holy grail of IIoT, itself poses many questions worth the trouble.

Trickle-in-the-Tub

The idea of people, machines and data intersecting on an industrial scale is undoubtedly huge but extrapolating the ‘data’ ROI of an IoT world to an IIoT world is not as easy or as impactful as it is may sound.

For industrial data, is different from individual device data. There is scale, complexity and a contextual flavor here that changes the wheels for IoT here.

Chadha avers to this complexity and relevance issue and adds factory integration to this mix. “The impact in an industrial environment is different.”

Kurlekar affirms that while the role of data remains the same - to provide more insights, however, the data belongs to and is owned by the industrial factory where the solution is deployed.

“The data cannot be used for marketing/ad/monetization purposes by other entities so the IIoT solution provider cannot make money off this data. It means inherently the costs of running and maintaining the IIoT solution needs to be recovered from the industry which has deployed the solution. This makes the operations expensive and recurring support, maintenance and service continuity costs are higher. Now if you factor the costs of these solutions running for decades, ROI is a real concern for both the customer deploying the solution and the vendor providing the solution.”

Does all this mean that upstream data flow over downstream flow will shift the question? In Chadha’s reckoning, it is all about dicing and slicing information at the end of the day and what matters is real-time information where it matters for the customer.

But then the security risk of upstream data flow is greater than downstream possibly. The security capabilities of an end node device such as a switch or a router are more primitive than that of controllers (cloud/non-cloud based) since the nodes typically have only basic factory set security that are standard across the board with less room for customization, Suresh explains.

So, of course, no matter how much we hate to say the 'S' word again and again, there is another dimension to all this data. The bigger question of Security.

Threats

All that sensory fabric and IoT connectors threading in industrial elements, hitherto scattered across miles, into one dashboard pops the lid of that much-dreaded and inevitable concern open: Doesn’t IIoT translate into a bigger, much easier-to-hack surface area for criminals?

If IIoT’s basic framework is being understood well, it can be readily surmised that for a hacker now sneaking into a vast industrial shopfloor would mean access to just one sensor in some corner.

The latest study from Honeywell Process Solutions (HPS) clearly indicated that as many as 92 per cent adults in India are fearful that cyber hackers are carrying out attacks on major industries and sectors of the economy in their countries.

There has to be something that is making industrial sectors likely to be perceived as vulnerable to such attacks. Look at oil and gas production (64 per cent), medical/health care/pharmaceuticals (64 per cent), power grid (63 per cent), chemicals (61 per cent) and aerospace/defense (59 per cent) and we can see how high the fears of a high-level industrial cyber threat are.

Kurlekar agrees that if the IIoT solution is breached or hacked it can be a real nightmare to industries. “Imagine if a hacker gets control over the machines in a plant, it can have disastrous results and with the kind of hacking/breaches that happen frequently, this is a very real threat to adoption of IIoT.”

“Increase in IIoT means more end nodes that are connected to the system on a two way basis which gives intruders the ability to hack into the system more. These end devices security protocols are not as secure as controllers. Eg:- layer 3 device is more secure than a layer 2 device (Routers and Switches).” Suresh chimes in.

IIoT systems can present an increased surface area for hackers. That is possibly why 37 per cent of those at enterprises are concerned about security when it comes to deploying machine-to-machine (M2M)/IoT technologies, Gillett’s analysis pointed.

Increase in IIoT means more end nodes on a two way basis, or more room for intruders: Suresh

Analysts hence strongly recommend enterprises to ensure that a Chief Information Security Officer (CISO) is on the case.

Asking for an inventory of IoT devices connected via IP as well as via nontraditional connection methods or working with the CISO to focus on security assessment, risk quantification, and eventually risk mitigation for each IoT item in the inventory can be a good idea then.

Tussle vs. Thrust

IIoT entails a wide variety of sensors, devices, radio technologies, and network and software protocols, plus high volumes of less common data types like time-series data, so as per Forrester the challenges of integrating IoT data and actuators into enterprise systems require new skills and technologies, too.

In essence, you have to treat IIoT as a service with multiple touch-points including sensors, gateways, OEMs, network infrastructure and its providers, cloud Infra and its providers, end points/mobiles, system integrations with other software/ERPs/DBs and more.

Different customers need different solutions: Chadha

Kurlekar insinuates that various components are developed, supported and serviced by various vendors and if any provider of these components stops support, the whole service can breakdown.

All these issues – a vendor's lifecycle clarity, interoperability, data’s complexity, data’s vulnerability or the very pragmatic struggle on collecting data or spreading IIoT in an industrial set-up, however do not and should not discourage the many positive implications that IIoT promises.

In the latest survey of the most innovative companies by The Boston Consulting Group what jumps out also is that 42 per cent of innovators reporting a culture of speed rated their organization’s innovation capabilities as strong, compared with less than 10 per cent of slow innovators. Long development times have been pointed as the most cited obstacle (by 42 per cent) to organizations’ ability to earn a return on its innovation investment.

IIoT can play quite a leveraging strength in turning tables on long development cycles. Forrester also has laid out how four new classes of enterprise IoT technology will emerge in 2016. In fact, for many products and markets, the continuous flow of data from IoT devices could be soon enabling a new “product-as as-service” business model. Understandably, IoT is spurring manufacturers into a new world of product hybrids (as pointed by Accenture too) altogether.

Innovation with IIoT can surely bolster new revenue models, speed and lean manufacturing for industries but if it can be achieved in an atmosphere where security, data’s use, infrastructure etc do not spill around as after-thoughts, then IIoT can be just the T that the industry has been waiting for – The Tour de Force.