Early in the third quarter of 2019, Gartner published its first-ever report on the Industrial Internet of Things (IIoT) niche which provided a picture into the IIoT market, cloud computing, edge computing, and their industrial applications. While the Gartner Magic Quadrant for IIoT showed an increased adoption rate, none of the listed stakeholders scored past the halfway mark in terms of execution. The execution score highlights the ability to apply IIoT data in resolving everyday problems across shop floors and every IIoT service provider was found wanting.
According to the report, “the installed base of IT/OT integration is small and focused on narrowly defined outcomes”. This means the integration of IIoT within shop floors isn’t as expansive as expected, and data analytics is still being limited to batch processing techniques which are far from the real-time analytics that Industrie 4.0 expects. It also highlights the fact that while cloud computing is increasing in adoption rate, it alone cannot deliver real-time IIoT data analytics and management.
This article will discuss:
- The challenges cloud computing faces in delivering real-time IIoT data analytics
- How real-time IIoT data analytics can be achieved and the role of edge computing and edge hardware
- The benefits of achieving real-time IIoT data analytics within shop floors
Challenges cloud computing faces with real-time IIoT data analytics
Although the extended computing resources that cloud computing provides have been advantageous to manufacturers, specific challenges exist when attempting to deliver real-time services. These challenges include:
Communication and network challenges – Networks are a source of unpredictability where real-time systems or data analytics is required. This means the transfer of data from IIoT devices to the cloud is subject to signal losses, noise effects, bandwidth limitations, and, in some cases, dynamic interferences from other systems within the shop floor. Although protocols that address these issues exist, real-time networking is still very unpredictable when only cloud computing resources are used to manage data transfer processes.
Distributed resource management issues – With multiple IIoT devices operating within a shop floor, the need to distribute adequate cloud computing resources to each device affects real-time data analytics. Although real-time scheduling algorithms can help balance resource distribution across multiple IIoT devices, these algorithms provide temporary guarantees as the operating systems within these devices makes it difficult to accurately allocate computing resources in real-time, thus affecting the quality of service (QoS) a centralized cloud computing environment can provide.
Storage cost and access control – To ensure the ROI of integrating cloud computing into industrial settings surpasses capital expenditure, enterprises will always have to watch their storage spending. In a situation where hundreds of IIoT devices, sensors, and systems are deployed, storage expenditure increases with increased data capture requirements. Thus, businesses have to find ways to ensure only important data is sent to centralized cloud ecosystems; and integrating edge computing is one way to achieve that.
Enterprises using cloud services provided by any vendor can also be troubled by downtime issues, which limit access to the cloud. When this occurs, real-time data exchange and analytics come to a halt and this disrupts any shop floor activity including IIoT devices that require the cloud to function properly.
Cybersecurity challenges – According to Infosecurity, over 100 million different attacks on IoT devices were detected in the first half of 2019 and the IIoT niche is no different. The increase in cyberattacks on IIoT devices is one of the biggest challenges with relying on cloud computing to solely handle data management activities. A successful breach will not only affect the hacked IIoT device but also provides loopholes that can be exploited to access an enterprise’s entire cloud-based infrastructure. Thus, data loss, phishing attacks, and ransomware affect the ability of cloud computing to deliver real-time data analytics.
Complementing cloud computing efforts with edge hardware and edge computing
At the recently concluded AWS re:Invent conference, one of the notable announcements was centered around providing supporting solutions to aid IIoT and cloud integration in industrial settings. This highlighted cloud computing’s need for supportive technologies if real-time management of IIoT data is to be achieved. Here, the integration of edge computing and the corresponding hardware that drives it is one example of complementary technologies that aid real-time data capture, management, and analytics.
With edge computing, many of the challenges cloud computing faces with delivering real-time data analytics are eliminated or drastically reduced. For one, edge computing supports real-time communication for IIoT devices (under 10ms) as it offers a physical middleware that allows IIoT applications to function in both an isolated mode and also facilitates data transfer to a centralized cloud environment. This means edge computing solves the problem of unpredictable communication and network challenges coupled with resource management issues.
The ability to capture data produced by IIoT devices and execute data analytics without accessing the cloud facilitates a more predictable process for managing IIoT data. It also eliminates the effects of interference, signal loss, and resource allocation when multiple IIoT devices and systems run simultaneously.
Edge computing also eliminates the need to constantly scale up cloud computing and storage resources with increasing workloads. IIoT systems can rely on an edge network to capture only the data relevant to a process and use it to automate tasks and develop operational schedules in real-time. Loss of stable connections and the issue of vendor downtimes will not affect real-time operations. This is because the edge hardware handles all the computing requirements of the IIoT system, and captured data can be transferred to the cloud any time communication is restored.
The ability to handle IIoT computing and data-related tasks outside of a centralized network enhances real-time analytics and keeps shop floor data secure if a breach to an enterprise’s unified network occurs. This reduces the effect of successful cyberattacks on the entire IT infrastructure and business operations of manufacturers.
Conclusion
The stringent timing requirements and high-level of predictability needed to deliver a functional real-time IIoT environment requires the use of more technological solutions than only cloud computing. Merging cloud computing with edge computing will go a long way to solve the complex problems associated with achieving real-time data management in IIoT-powered environments. This will ensure cloud computing can then be applied to solve difficult real-time problems to advance the implementation of Industrie 4.0 models.
