Skip navigation
4 Things You Must Know about the IIoT

4 Things You Must Know about the IIoT

Understanding current trends, security, and future perspectives in the Industrial Internet of Things will benefit any sized company.

With new data services, online apps and dashboards can be accessed for remote use via computer, tablet, or phone. A data service company might charge depending on how much access or which features a client wants to use.

The Industrial Internet of Things (IIoT) is slowly transitioning from being a mere buzzword to a concrete reality. Innovation in this field is mainly led by large companies that have already started to implement IIoT ideas in various applications. Now they are able to observe the results of their investments in IIoT.

Some of the big players in the IT world believe that competition, resource constraints, and an aging work force are just a few of the key elements that are driving innovation toward IIoT. A company that decides to invest in this technology will observe a clear advantage over other competitors in a relatively short time. The availability of a new set of analytics, generated by continuous monitoring of real-time data, will enable that same company to gain a deeper insight into its production process, and also to reduce costs by reducing waste, downtime, and unnecessary maintenance.

According to a 2013 report by Aberdeen Group, 53% of U.S. manufacturers that implement IIoT have improved their business, increased their competitive edge, and reduced total costs. For 40% of U.S. manufacturers, the biggest risk factor is the failure of critical assets. Whereas the impact of certain components is shrinking, such as logistics and supplier quality, some other areas are becoming crucial for the success of a company and need more efficient analysis tools.

1. Acquire Smarter Talent and Maintenance

Product failure is still the top risk component, but far more interesting is the effect of a new factor that emerges from Aberdeen Group’s study. “Failure to acquire and retain talent is becoming a real challenge for every company that aims to be a market leader. In our opinion, this is fundamental. We are enabling our machines to be smarter, but in order to do that, companies are required to have a more technically skilled work force that understands new technologies and that keeps itself constantly up to date.”

Today’s production machines are complex manufacturing systems that are dynamic and coupled, and the environment in which they operate is highly dynamic and highly coupled. IIoT allows us to cope with this demand of highly dynamic and highly coupled systems, moving from monolithic, slow-paced systems, to fast, on-demand modular systems. One example to demonstrate this shift is the predictive maintenance (PdM) model. Until now, PdM has been implemented using mean time between failures, often called MTBF. This time-to-failure model is stochastic, generalist, and therefore not accurate. The time frame of reference in this case is monthly-based maintenance. Downtimes are planned based on a machine’s operating time and sometimes machine parts are replaced even if they were not broken and still fully functioning. With IIoT, the paradigm of maintenance shifts from predictive to reactive. Monitoring every machine using its own particular operating conditions means scheduling downtimes can happen days before a part is supposed to break. Manufacturers can react to data generated by the machine. In some cases, it is also possible to predict precisely how many operational cycles are left before a breakdown.

2. Understand the Importance of Big Data

IoT is a network of connected smart devices, able to communicate with each other, but also able to send information to a storage system, local or cloud-based, to analyze and refine data to gain better context-related knowledge. The key is in the data, and many believe that having more data is better than having better data-mining algorithms.

It is important to understand that, in general, IoT is a Big Data problem, and for this reason we need Big Data tools to analyze it. Big Data is essentially composed of three elements:

Volumes: Massive amounts of information are produced…

Velocity: … in a very fast manner and with unprecedented frequency.

Variety: Data is generated by multiple sources, containing heterogeneous information, and in the majority of cases, not structured.

In a scenario where human operators work with an IoT system, the operators are considered the only bottleneck. A 2013 ABI research report stated that by 2020, there will be 30 billion connected devices. In the same year a Morgan Stanley report stated that by 2020, there will be 75 billion devices. Intel believes that in the same year the number of devices will be 31 billion. The results coming out of market research and predictions might not coincide 100%, but the main point is there will be a massive increase of new IoT applications, a huge demand for bandwidth, connectivity to cloud solutions, and secure access to data analytics.

Considering an average of 50 billion connected devices by the year 2020, we can estimate that the necessary bandwidth to cover all of the IoT needs will be about 20 million TB per month, an enormous demand! Dealing with such large amounts of information can be challenging. The good news is there are actions to take to alleviate this burden.

The IoT mantra says that all data must be collected (really, all of it!) to be then analyzed in a different place to extrapolate a new level of knowledge that is more useful to the data owner. A possible downside to this is the creation of huge “data museums,” where raw data is just sitting in some kind of cloud storage, losing its value waiting to be used. A much better approach would be to collect all data and then generate some “structured data” by performing some simple processing right where they originated (i.e., sensors, gateways, or data collectors). This is referred to as “fog computing” or edge computing.

Looking at IIoT from an end-user perspective, we can observe how this technology has the potential to enhance production processes. IIoT is the convergence point between information technology (IT) and operating technology (OT). These are typically two very distinct departments, but IIoT forces them to interact and cooperate to understand each other’s needs. IIoT can contribute to making machines smart by adopting the newest connection protocols to allow machine-to-machine interaction, but most importantly, by giving more timely and detailed information to external operators. Sometimes, IIoT has pushed manufacturers to invest in retrofitting old equipment or even convinced them to buy new machines. For all of these reasons, IIoT aims to improve efficiency, reliability, and productivity of operations, with a noticeable reduction in cost and waste.

Bandwidth and local processing will help ease the massive amount of data that service providers and cloud resources will have to handle as more automated equipment is connected over the internet.

3. Utilize Service Models

On the other hand, if we look at IIoT from an OEM perspective, we can observe that this is not just a technological revolution but also an economical one. In fact, it allows the use of completely new business models, which could not exist otherwise. For example, imagine a machine builder starts to implement a network of IoT sensors on board of each machine produced. Those sensors can collect information on time, air pressure, temperature, scrap, work pieces, lot number, together with some more canonical information on alerts, errors, and downtimes. Customers that are interested in optimizing and reducing costs might want to have all of this on each machine in their production line, and therefore they could pay more for this service. All data generated belongs to the client (this is a commonly accepted specification), but an OEM could also provide all the infrastructure and connectivity to store data. Whenever a client needs more refined and structured analytics, the OEM can add them to the client’s dashboard, or web services, with apps, or through other systems for data interaction. We could also provide a rule engine tool that allows the clients to create custom rules through which they can get notifications or warnings if certain operational conditions are not met. Finally, and this is where the IIoT really shines, we could gather better insights on our clients’ machines’ state (anonymously or on their behalf) and predict with accuracy when a certain component will break, allowing OEMs to send spare parts before the breakdown happens unexpectedly.

The advantage of this approach is evident. With the OEM generating new models for improving customer satisfaction and generating revenues, and with an accurate cloud-based service, end users can count on reducing unexpected downtimes. Even the smallest family-owned company can benefit from these services to gain a competitive edge over other companies that do not use their resources as efficiently.

The final perspective is given by those agents who create and maintain communication and cloud infrastructures. By 2020, only 5% of those 50 billion devices that we estimated previously will be connected through a cellular network, mainly those displaced in remote locations or with particular shape factors. All the rest will leverage more canonical connectivity, via routers, access points, or gateways specifically designed for IoT. The cost of storage is going down, but maintaining a server implies high costs, due to electricity, cooling systems, operators, and other miscellaneous factors. Nowadays there are many big companies providing cloud services on large scale, which are easy to use and relatively cheap for end clients.

Thanks to IoT, those big companies have been able to adopt new business models and to provide services that are more suitable for smart devices. For example, Amazon is shifting its cloud-service paradigm from storage-based to computing-based. Clients can decide where to store their data, and independently from their location, a client could then use a specific function (a piece of code) running somewhere in the cloud on his own data using a set of application program interfaces, known as APIs. The cloud provider at this point can charge for each function execution. Somewhat similar to this is the road taken by IBM, which is revolving all of its cloud platform around an IoT cognitive advisor—an artificial intelligence called Watson. A client is charged to have his or her data analyzed by this virtual brain, which takes some structured data as input and gives back output predictive analysis. This type of service can make recommendations on how to optimize the process and how to avoid factors that lead to quality problems. All investment efforts in this case reside almost exclusively in the formulation of a better, smarter, self-learning artificial intelligence. Other popular cloud-based solutions include Bosch IoT Cloud and Oracle Cloud.

4. Avoid These Pitfalls

Having all the data from all the machines of a certain company traveling over the internet is not necessarily a reassuring scenario, especially if those machines control critical processes, such as nuclear plants, water purification, or hazardous material production lines.  In such situations, these concerns are well founded. A quick search on the internet can lead to a particular search engine designed for IoT and smart devices ( Anyone, from anywhere in the world, can scan for certain IP addresses, find out their location, what type of device they are associated with, and what OS is running on them. With the right skills, and some luck, it would be possible to generate destructive consequences with just a few mouse clicks.

Security is always a concern for IIoT. One way companies are keep systems safe is to segment production with protective means, such as firewalls and encryption to offer access to employees, outside vendors, and services while maintaining security.

Security is not always considered as carefully as it should be in our day-to-day applications. These problems are not new; they didn’t just arrive with the implementation of IIoT devices.  But they do show that we are not used to thinking in a secure way when dealing with our internet-connected devices. Sometimes users might be failing to deploy even the most basic security measures in order to guarantee a minimum level of defense. The advent of IoT is simply forcing us, now, to think about the things that have become the most valuable to us—data.

If we imagine an IoT infrastructure as a connected graph, a malicious attack could come from any point in the net, from any node and from any edge. The easiest is the so-called man-in-the-middle attack. The attacker intercepts a connection, forces a disconnection, then grabs the packets that the client sends in order to reconnect. The attacker then has access to the user’s credentials and could use them to establish a valid connection.

Another simple attack is DoS, Denial of Service. This type of attack was made public most recently on Oct. 21 when a major digital naming system hosting service with registration was attacked, resulting in major outages on such sites as Twitter, Paypal, The New York Times online, and others. The attacker floods the server with a huge amount of data in a brief time span, causing the server to crash. There is also code injection. This typically works with web apps. The attacker uses an input field (even a comment form on a website might work) to input some JavaScript code that has the function of modifying the front-end code, so that he can grant access to methods, functions, or data that were not intended to be exposed to the public.

All the major players in the IIoT world seem to agree on a few preliminary steps that anyone can take to increase the security of a network at every level. First and most important, think like an attacker—for instance, try to hack your own system. Look for weak points and try to gain access to private information. Identify vulnerabilities, both at the software and hardware level. Validate your design, keeping in mind security principles and identify countermeasures that can be added. Sit down with your engineering team and come up with a Risk Management Assessment, scanning every node and edge on your network identifying all the possible attacks. Focus on robustness. Include several security layers in your design and think about how to react to certain attacks and how to recover from them.

Remember, an attack can come both from hardware and software. Software security is a well-known topic, but hardware security is not always taken into account. Some hardware manufacturers are already designing innovative chips that will hold security keys installing them directly on the piece of equipment, making it impossible to tamper with, but also allowing a faster and more reliable software authentication. Risk cannot be eliminated; it can only be managed.

IIoT is not just a niche in the hands of a few technology firms dealing with IT and software as a core part of their business. On the contrary, IIoT has made its way into various and heterogeneous fields. I have seen applications from the transportation sector, industrial automation, oil and gas, pharmaceutical, energy distribution, even food and beverage and agricultural.

The strength of IIoT is to enable the monitoring of a wide range of variables in a small amount of time, tasks that the human brain is not suitable. Small intelligent sensors can combine their data in order to create a higher layer of information that is accurate, fast, and otherwise unobtainable.

Data is around us, whether we like it or not. Deciding to harness its power is not just a technological innovation, but it is quickly becoming a need for companies to remain competitive in the marketplace.

This is just the beginning of this new smart-device revolution and there is still a long road ahead. Being part of it is not just exciting, but also a vital part of the future of every company. 

​Matteo Dariol is a control system engineer focused on IoT and Indystry4.0 technologies. He received his B.E. from L’Aquila University (Italy) and his M.E. from La Sapienza”University of Rome (Italy), majoring in adaptive control strategies, robotics, and machine learning. Currently he is a product developer for Bosch Rexroth, but his professional experience ranges from embedded systems development, to R&D, automation, web front-end and MES. Before arriving in the U.S. for a Ph.D. program in aerospace engineering, he lived across various European countries contributing to multidisciplinary academic projects.


Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.