IIoT arrives; It’s time to get started
Implementation is not only practical, but now is a competitive necessity.
The Industrial Internet of Things (IIoT) has progressed from dream to hype to reality. Today, the basic deployment scenarios of the IIoT solutions we implement for our end user manufacturing customers include:
- Greenfield deployments, which are primarily found in "smart" solutions related to advanced monitoring and visibility
- Brownfield upgrades, which are the introduction of IIoT technologies and approaches to existing facilities to expand asset and process visibility and analytics
- New asset-monitoring services from vendors who are leveraging IIoT to provide remote predictive analytics capabilities for their assets installed at customer sites.
The enthusiasm and growth in IIoT momentum over the last year is obvious in the pages of this magazine, as there continue to be IIoT-related articles in every issue. But this does not mean the industry has agreed on its naming conventions; so we continue to refer to "digital transformation," "smart manufacturing," the "fourth industrial revolution," and "Industrie 4.0", in addition to IIoT.
Regardless of what a sensored, connected, integrated factory is called, IIoT continues its march forward as the expected architecture for process manufacturing facilities. The competitive advantages enabled by increased and consistent visibility, accuracy and data-driven insights on production results are simply too important for most firms to ignore, hence the growth in IIoT deployments.
Don’t get caught behind
As we look ahead in 2017, there are many ways to discuss what’s ahead for IIoT. For example there are new IIoT technologies to consider, such as drones, robots, voice-powered artificial intelligence (AI) and virtual reality solutions. And while those will likely have a future in IIoT, there is also an important issue we need to address first: most firms need to catch up to the technology opportunities available today.
Analyst research shows a large gap between the expectations and advantages of IIoT, and the state of current deployment efforts. Further, IIoT deployments are generally piecemeal rather than broadly applied across a plant. What this means is that IIoT enthusiasm and potential benefits are not being realized by many end users, with the opportunity still ahead of them, instead of being already recognized in bottom line results.
Part of this gap may be due to the oft-cited advice that companies should start small and have IIoT deployments prove their value and impact before moving on to more widespread implementations. This is great advice and hard to argue with, and aligns with the limited number of large IIoT deployments within companies. Certainly an early success will bode well for organizational enthusiasm for continued IIoT investments. End users should be complimented for any IIoT investments or trial deployments they have completed, because at the end of the day they have done something and at least tried. In the world of software startups there is an expression: "fail fast." This means it’s better to try something and learn it doesn’t work than to wonder if it might have worked. So any end user with something to share, prove or present with respect to IIoT is ahead of those still wondering what’s possible. The rest of this article will examine the challenges and experiences of end users who have taken the leap, with the goal of providing a roadmap to others contemplating similar steps.
Standards, security are issues
The quip about standards, "the great thing about standards is there is one for every issue," continues to be a reality in the IIoT ecosystem. Perhaps 2016 was a step in the right direction with partnership agreements and the merger of some standards bodies, or maybe 2016 was a step in the wrong direction with ever more standards and the momentum of competing efforts. Either way, 2017 will not see a world of functional, compatible, Lego blocks which an end user can assemble to create a mixed-vendor, distributed, IIoT solution.
Meanwhile the other "s" word-security-is consistently a leader on lists of requirements, concerns and issues with IIoT-and certainly in this area will not be realized in 2017. Automobile, HVAC, SCADA and any number of other systems have been hacked. Unfortunately, there is no easy answer to security issues. Instead, what is needed is hard work, best efforts and the selection of solutions with limited exposure to the outside world.
Using either a lack of standards or fear of security issues as an excuse to not move ahead on IIoT implementations simply pushes the necessary learning and experience that will come out of the effort further down the road,and can put end users further behind in the IIoT race.
Focus on business impact
In many IIoT articles, the starting point is the sensor. That is the origination point for data that flows through a communications network to a centralized application with analytics, integration, data storage, etc. It is easy to start at the sensor because the hardware and communications innovation that is driving IIoT is exciting and it is fun to think it about from that perspective.
Micro-sensors; Arduino, Raspberry PI and Intel Galileo platforms; long-life batteries; low power wireless systems; and builder kits from Microsoft, IBM and Amazon are areas in the news and of interest. There is so much to imagine with sensors and logic added to every possible asset and structure.
Unfortunately, these technologies often don’t necessarily serve the business end of the issue. The right question is not what’s possible with the new technologies, but what’s important to the business: a quantifiable positive impact to production and business outcomes.
This impact can be realized in many ways, such as increased uptime, improved quality, higher yields, etc. Positive results also could include motivating and incenting employee behavior through visibility to production status and the impact of their actions.
The best IIoT deployment stories are those demonstrating positive impacts to the bottom line. These stories stress the impacts, then work backwards to the technologies making benefits possible.
How do you add value?
Remote monitoring services represents the transition from an asset view to a capability view. This opportunity is typically framed from the point of view of the asset vendor, but probably should instead be framed in terms of end user benefits. GE, for example, talks about moving from selling turbines to selling services.
These new services, powered by IIoT infrastructures, represent an opportunity for asset vendors to increase their revenue, but more importantly to add value to their end user customers. From an end user’s point of view, the idea of additional revenue for asset vendors from smart connected products is not a likely source of interest. But it does raise important questions: Who is best suited to extract value from asset data? Who is best positioned to monitor, analyze and recommend outcomes on asset performance, the vendor or the end user?
Companies routinely outsource services to specialized organizations, from cafeterias to accounting to maintenance. But does this model make sense for IIoT deployments? The value of reduced scheduled maintenance alone could justify the cost of an outsourced monitoring service in some cases, so the answer would be yes in these instances. But in other cases, remote monitoring services can be a disruptive issue for current employees and processes, so care must be taken when deciding what to outsource and what to keep in house.
By tapping expertise on assets from vendors, end users can focus more on results from the value created by the integration of assets than on the status of a single link in the process. If the end user company’s differentiating expertise is in asset optimization excellence then this may be of less importance, but the focus on where an organization creates the most value is a question to ask.
IIoT solutions typically rely on the assumption that at some point the "magic happens here" to close the gap between data and insight. Typically this magic is buried under a banner of machine learning, big data or advanced analytics. But just where these key insights are revealed, and how, is often given insufficient consideration.
Looking at end user examples, the real work of analytics includes data aggregation, cleansing and contextualization with business systems data before insights can be produced. Or as our end user customers describe it, 80% to 90% of analytics is getting the data right before the analytics can occur, and this will only get more complicated as end users install more connected products which will provide even more streams of data for integration (Figure 1).
There are also issues of engineer productivity, team collaboration and process industry capabilities to recognize in the context of analytics offerings. These are hard issues, and therefore the specifics of analytics requirements deserve the same attention as the business cases.
We certainly know from experience there are no silver bullet solutions to actionable insights. Our experience at Seeq suggests the right path is instead providing end user engineers and other production experts with tools they can use to create actionable information from raw data. These tools rely heavily on visualization of data, from which insights can be derived (Figure 2).
Finally, to land the analytics once insights are achieved, it’s important to know which employees are trained and incented to act on the information that has been uncovered. A "right" answer disconnected from the incentives and rewards of the workforce is a recipe for failure. Therefore, identifying the requirement to find the insight, executing the analytics and transforming insight into execution is the last mile of successful IIoT implementations.
Getting started with IIoT
Any new project or proposal requires effort to overcome the inertia of doing nothing. IIoT projects are in the balance between the availability of necessary technology and the risks and rewards of execution, a difficult juncture.
The issues listed in this article are not the only ones needed to make IIoT real, but they often have been the main stumbling blocks, and the main reasons why inertia hasn’t been overcome to start an IIoT project, or why projects fail to land the expected impact. Getting started by investing in analytics and expending the required effort to deliver business results on IIoT can be the hardest part of progress, but it’s worth it for many end users. For those already engaged, a strong congratulations, and for those waiting in the wings, it’s time to succeed with IIoT.