Insight

Unifying IoT across the Digital Enterprise

The path forward

Bret Birnbaum

Bret Birnbaum

Director Advisory, Lighthouse, KPMG US

+1 347-526-1745

The promise of the Internet of Things (IoT) revolution has finally come to fruition. A wide spectrum of products, tools, and infrastructure components is being equipped with sensors, creating a global network of “smart” and connected devices. As this network grows and diversifies, so do the benefits of living in a connected world. Businesses can unlock new markets and insights, leading to leaps in efficiency and optimization not previously possible.


Challenges and opportunities

While the future of IoT is bright, leveraging an enormous amount of new data comes with its share of challenges.

Innovators are realizing that simply connecting smart products to the internet does not make them more valuable. True value comes from capturing, storing, analyzing, and visualizing IoT data to derive meaning. To this end, certain industries are demonstrating real progress. For example, medical device manufacturers have begun to create cloud platforms that can aggregate sensor data from patients and present it back via web applications, mobile apps, and analytical tools. These tools then analyze the data as the basis for self-care recommendations, changes in treatment protocols, and insight into disease progression. Moreover, manufacturers of these devices can use predictive analytics to monitor the health of the devices themselves to ensure they remain in optimal working condition for patients and predict failure before it actually occurs.

IoT solutions offered today, such as those used on the factory floor in manufacturing, are primarily focused on devices and data alone. This is problematic because siloed data and single-purpose applications do not help digital enterprises harmonize workflows, increase information access, and feed into user applications across a firm. Having specific IoT applications for specific devices makes it very difficult to solve problems across the firm. Users tend to become frustrated when they need to use a multitude of applications to complete their work, not to mention cutting and pasting data from one application to another. At best, this is a disjointed and inefficient experience; at worst, the processes are not scalable or measurable. 

The issue of consistency across the enterprise can be illuminated by looking at heavy construction equipment manufacturing. This industry has done an excellent job equipping their machines with tightly integrated sensors that provide insight into equipment health and performance, unlocking massive value in predictive maintenance, usage monitoring, and more. Further, many organizations have enhanced their offerings with robust cloud platforms that can aggregate machine data in an environment where users interact and analyze. 

While this strategy makes sense from the equipment manufacturer’s perspective, enterprises that buy their products rarely source from one supplier. For example, if a company buys bulldozers from two suppliers, employees have to swivel between two different IoT platforms to see data for the entire fleet. Since those systems are proprietary and separate, it is difficult to build business logic across their fleet as the data is siloed and ROI is ultimately reduced.

As the number of IoT offerings continues to grow, the cost and effort to bring on new solutions become prohibitive and unscalable. Many enterprises are eager to implement new IoT use cases but become frustrated when they find out they need a whole new set of applications and data stores unique to each implementation. Maintaining, customizing, and integrating multiple platforms is not only cost prohibitive but overly taxing on an IT workforce that is often already stretched thin. For example, if HVAC data/control is siloed in one application and space monitoring in another, then the ability to turn down the air conditioning when a room has not been occupied for an extended period becomes a complex IT integration project. The IT staff need to understand the data structures and development technologies which could be vastly different requiring different skills. Moreover, the integration between these two systems requires workflow and API technologies. The additional complexity adds time, cost, and increased chance of error.

Homing in on actionable insights

Innovators are realizing that simply connecting smart products to the internet does not make them more valuable.  True value comes from capturing, storing, analyzing, and visualizing IoT data to derive meaning.  To this end, certain industries are demonstrating real progress.  For example, medical device manufacturers have begun to create cloud platforms that can aggregate sensor data from patients and present it back via web applications, mobile apps, and analytical tools.  These tools then analyze the data as the basis for self-care recommendations, changes in treatment protocols, and insight into disease progression.  Moreover, manufacturers of these devices can use predictive analytics to monitor the health of the devices themselves to ensure they remain in optimal working condition for patients and predict failure before it actually occurs.

Integrating siloed systems and data

IoT solutions offered today, such as those used on the factory floor in manufacturing, are primarily focused on devices and data alone. This is problematic because siloed data and single-purpose applications do not help digital enterprises harmonize workflows, increase information access, and feed into user applications across a firm. Having specific IoT applications for specific devices makes it very difficult to solve problems across the firm. Users tend to become frustrated when they need to use a multitude of applications to complete their work, not to mention cutting and pasting data from one application to another. At best, this is a disjointed and inefficient experience; at worst, the processes are not scalable or measurable. 

The issue of consistency across the enterprise can be illuminated by looking at heavy construction equipment manufacturing. This industry has done an excellent job equipping their machines with tightly integrated sensors that provide insight into equipment health and performance, unlocking massive value in predictive maintenance, usage monitoring, and more. Further, many organizations have enhanced their offerings with robust cloud platforms that can aggregate machine data in an environment where users interact and analyze. 

While this strategy makes sense from the equipment manufacturer’s perspective, enterprises that buy their products rarely source from one supplier. For example, if a company buys bulldozers from two suppliers, employees have to swivel between two different IoT platforms to see data for the entire fleet. Since those systems are proprietary and separate, it is difficult to build business logic across their fleet as the data is siloed and ROI is ultimately reduced.

Overcoming resource limitations

As the number of IoT offerings continues to grow, the cost and effort to bring on new solutions become prohibitive and unscalable.   Many enterprises are eager to implement new IoT use cases but become frustrated when they find out they need a whole new set of applications and data stores unique to each implementation.  Maintaining, customizing, and integrating multiple platforms is not only cost prohibitive but overly taxing on an IT workforce that is often already stretched thin.  For example, if HVAC data/control is siloed in one application and space monitoring in another, then the ability to turn down the air conditioning when a room has not been occupied for an extended period becomes a complex IT integration project.  The IT staff need to understand the data structures and development technologies which could be vastly different requiring different skills.  Moreover, the integration between these two systems requires workflow and API technologies.  The additional complexity adds time, cost, and increased chance of error.  

The path forward

Many enterprises envision a digital future using IoT, but are frustrated by vendor lock-in and siloed use cases that prevent them from scaling their investments. Luckily, as with most technical problems, there is a technical solution.
 

    

The “Enterprise IoT Platform”

Enterprises can decouple themselves from multiple vendor IoT platforms by implementing their own master IoT platform that captures sensor data from across the enterprise. Access to funding may be given a boost if IT teams can present senior management with enterprise-wide, business-driven use cases regardless of the number of data types and sources.

Moreover, a single data model can be used for data attributed across multiple vendors. For example, in the case of the two bulldozer suppliers, one vendor may use gallons for fuel levels and the other liters. Having a single data store allows data to be transformed on collection to the units used by the enterprise, creating an apples-to-apples comparison.

Further, companies in this position can achieve data consistency with the use of distributed computing. In a distributed model, IoT sensors are technically independent of the location where they report data. They can be programmed to report directly master IoT platform, the vendor’s platform, or both. In the latter case, cloud-to-cloud integrations are used to pull data from the source systems in nearly real time.

    

Direct communication with the platform

IoT sensors are programmed to send their data “home,” which most often means the cloud.[1] Although some device suppliers don’t have a platform, when they do, it is likely that customer data will be directed back to them. Alternatively, companies can work with their suppliers to update their device-sensor firmware to have them point data to the customer cloud platform rather than the vendor cloud platform. As another example, companies using cellular backhaul technology can purchase and manage their own SIM cards for their devices. These SIM cards can be programmed to send data back to the company cloud IOT platform. 

In both cases, data bypasses the vendor cloud all together. The main advantage here is access to real-time data directly from the devices themselves. It is important to bear in mind, however, that there are some drawbacks to this model, including the technical challenges of maintaining SIM cards and potential network costs without economies of scale.

   

Cloud-to-cloud Integration

For scenarios where device makers cannot change the destination of device data or where it is preferred to use the pre-installed sensors that report into their proprietary web cloud, cloud-to-cloud integrations are used to extract sensor data and stream it into the enterprise IoT platform on a regular basis. This model is quick to set up and easy to maintain. 

Cloud-to-cloud integrations are typically achieved with RESTful APIs. Calls are made to these APIs on a regular schedule to pull the latest IoT data for each device stored in the vendor’s cloud platform. Using the bulldozer example above, if the two vendors already have tightly integrated IoT sensors that report into their proprietary cloud platforms, the enterprise can periodically pull this data from each respective cloud, say every 15 minutes, to get the latest IoT data for each machine. IT teams can pick and choose select data elements to store in their platforms, clearing away copious amounts of unneeded data, which saves bandwidth and disk space.

Data flows from the machine to the machine manufacturer’s cloud platform to the customer platform.
Brett Birnbaum, Advisory Director, KPMG US

The primary benefit of this configuration is that the organization benefits from the manufacturer’s investment in IoT hardware and network connectivity. The second major benefit is that manufacturers gain access to data that allows them to gain detailed insights into the machines’ operations, which could help with maintenance and future product quality enhancements and features.

As an enterprise implements more and more IoT use cases it is almost guaranteed that both paradigms will be implemented. The takeaway point is that no matter which paradigm is used, the enterprise will have a single platform for all IoT data from which it can start building digital applications that span enterprise-wide use cases.

Footnotes

[1] Devices can also report into Edge computing devices. Edge computing is out of scope for this report.

Choosing the best platform fit for your organization

Deciding to implement an IoT platform to unify your IoT strategy is only half the battle. Choosing the best platform to meet your business needs and align with your company’s IT capabilities is critical to ensuring success. Luckily, there are numerous sophisticated IoT platforms from which to choose. 

Almost all of these platforms are available on prem or in the cloud (PaaS, SaaS). Some vendors offer a vast array of tool sets (e.g., Microsoft Azure IoT Hub, AWS IoT) that allow for a high degree of customization and integration, although they require deep enterprise-architecture and application-coding skills. Organizations that already use Cloud platforms, develop their own applications, and have access to a large pool of IT talent will be in the best position to benefit from these types of platforms.

Other solutions, such as PTC ThingWorx, provide end-to-end IoT platforms that accelerate the organization’s application deployment time and speed to market. These platforms come with tools and pre-defined models for setting up IoT devices, representing them as digital twins. They often come with low-code platforms that let the organization quickly design, build, test, and deploy applications to its internal and external customers. These types of platforms provide an excellent opportunity for organizations with a high degree of product and operational knowledge but limited IT capacity to enter rapidly into the IoT space, while forfeiting a small amount of control. Additionally, if the organization grows its IT skill base and wants to create more elaborate applications, it can forego the low-code aspect of the platform and use the connectivity and data platform while building custom applications that integrate.

Conclusion

In the past, technology initiatives tended to focus on specific business functions (e.g., CRM, WMS, etc.). Today, IoT use cases cross many business functions to provide connectivity and new value. Deciding which IoT use cases to implement across the enterprise requires business functions to work together towards a shared solution that fits within the overarching business vision of the company. If an integrated approach to IoT is successful, then value should be measurable from both a qualitative perspective (e.g., user acceptance) and a quantitative perspective (increased revenue/reduced costs). 

The internet of things is not a theoretical technology anymore. IoT and the advantages that come with it are here to stay. The question businesses are facing is not whether they should invest in IoT, but when they should invest. The answer is the sooner, the better. The data that drives the value of IoT accumulates over time. Therefore, every year that passes without acknowledging IoT is a missed opportunity for applying its insights to the growth of the company.

The reality for most businesses is that IoT data will come from not one, but many connected sources. The growing variety of IoT products available today means that innovation is progressing rapidly. Real success is possible if IoT product manufacturers move away from operating within their own digital platforms and implementing solutions without consideration of one another. In contrast, taking an integrated approach to deploying multiple IoT use cases across an enterprise allows each IoT solution to coexist within the same space. Most of the time, the best way to accomplish this is to route all IoT data sources into a unified IoT platform. By doing so, businesses can aggregate information, perform deep analytics on the data, and then make the insights gained accessible to all, in a single location. Enterprise IoT platforms such as PTC Thingworx can help implement an IoT ecosystem for the entire enterprise.

Contact us

Megh Desai

Megh Desai

Director, Advisory, Procurement & Outsourcing, KPMG US

+1 630-596-3325