top of page
blog-bg.jpg

Designing Industrial IoT for Elasticity and Resilience

The arms race for the adoption of the Industrial Internet of Things (IIoT) has entered a phase of exponential acceleration in 2020. Success in the implementation of IoT projects is more common as the industry emerges from the trough of despair. The industry now knows how to plan and design IoT projects, at the outset, to succeed. With many moving parts in IoT systems, design helps build resilient foundations to make them failure-proof.


Industrial Internet of Things gains momentum


VDC Research reported the turning point in Industrial IoT (IIoT) adoption in its research report “The Cost of Being Late to the IoT” completed in March 2020. It found that engineering organizations, 78% of them, expect IoT will have at least some impact on their business over the next three years— more than double of the current rate of 30%.


Among the compelling reasons for the faster pace of investment in IIoT is strategic opportunities for serving customers in new ways with services (25.6%) and competitive pressures (18.2%).  Predictive analytics, for maintaining the health of equipment and other assets, is a prominent service in the IIoT environment.


Benefits of implementing the IIoT


The companies that have adopted IIoT lowered their development time by ten percent as the likelihood of them embracing modern development methodologies, such as continuous integration, is higher by 50 percent and virtual prototyping three times more. As a result, they realize cost reductions of an average of over thirty-three thousand for each project according the VDC report.


One-third of manufacturers reported that they used data from their smart devices to improve their operating and manufacturing processes according to a PwC report of 2017.

Opportunities for earning revenue from new sources expand in the IIoT environment. From among those who have adopted IIoT, forty-seven percent reported they earn revenues from more IoT-enabled products, thirty-nine percent from selling data, and thirty-one percent bundle professional services with the products.


The demanding challenges of designing IIoT


Industrial IoT systems are a labyrinth where data is sourced from a heterogeneity of countless sensors and devices which use a diversity of formats. The data is not usable for analytical purposes without a consistent format to access all sources of data before it can be aggregated and analyzed.

Starting from a foundation of data collection technologies, intersecting IT, networking technology, cyber-physical technologies, and intelligence services transport, store, and analyze data at gateways or clouds. Software adds a layer that virtually interconnects them. The risk of failure at any of these junctions, where individual technologies interconnect, is high and can cascade to the overall system.


Troubleshooting for these systems is tortuous, if not impossible, because it is hard to diagnose the root cause of a problem in a thicket of a system after the fact. These systems need to be resilient to weather most adverse events, from the outset, to avoid the high costs of downtime.

Networks that are conduits for the flow of data within the premises of factories must be able to reach every nook and cranny of a factory for a comprehensive collection of data. Despite the numerous sensors and devices that are connected, the network costs must remain low for the investment to be economically viable.


Data volumes grow rapidly and oscillate as new analytical applications are written and their usage unexpectedly spikes. Networks need elasticity to meet the fluctuating demands within factories to avoid congestion or underutilization of their capacity. Networks that operate in dedicated frequencies and have fixed number of nodes are overwhelmed when data traffic spikes.

Some use cases have low tolerance, if at all, for any deviance from the expected metrics set in service-level agreements. Typically, these are software controls for sensitive equipment where low latencies and high reliability are imperative for receiving data and executing controls to prevent failure.


In the IIoT environment, networks need to be able to receive data from massive numbers of sensors with a high quality of service and outside the perimeter of a factory. A typical use case is OEMs monitoring the performance of the equipment they supply.


Designing the foundations of failure-proof scalability


Software-defined IIoT systems and networks are failure-proof when they are designed, from the ground up, for versatility and elastic scaling. They are versatile in gaining access to data from the gamut of heterogenous sensors, devices, and equipment, when they are interconnected with software and APIs.


These systems gain scalability by building on-demand capacity with software. The software-defined systems build a pyramid that includes hardware underpinning multiple hypervisors which serve several virtual machines that support a few operating systems. In turn, the operating systems are the backbones of multiple applications built, on-demand, by assembling microservices. Any of the software components can be used interchangeably depending on the volume of traffic and workloads.


Spikes in traffic have a low probability of overwhelming these systems because loads can be flexibly shifted when capacity falls short at any one facility or location.

Some of the solutions for specific challenges of versatility, scalability and resilience are as below:


  • Middleware: uses a common format to access data from heterogenous data collection software concurrently, and distributed data services acquire the data from middleware at multiple locations to classify it with metadata for analytics

  • Redundancy: virtual machines are spun in short order to create capacity proportionate to data volumes

  • Predicting failure: software systems are monitored constantly and the operating data flowing from them is used to predict failure and replace failing components before they break down

  • Remote maintenance: software systems provide visibility to the state of the systems and networks and corrective action is taken remotely with software controls and software and security updates

  • Network slicing: segments of networks are sliced and customized for pre-determined levels of service quality such as low latency and high reliability for software controls

  • Mesh networks: peer-to-peer networks that interconnect installed devices and machines bring data from sensors in every corner of a facility into wide area networks inexpensively

  • NB-IoT: is a last mile wide-area network that can access data from massive numbers of heterogenous sensors and devices


Conclusion: IIoT is promising for the future of the manufacturing industry but its deployment is not necessarily successful. The design of their foundations separates success from failure.

0 views0 comments

Σχόλια


bottom of page