It has been abundantly clear for quite some time that enterprise technology development has been focused on the digital workplace, but according to a recent book from Tom Seibel, which we discussed last month, what is happening now is a game changer.
In Digital Transformation: Survive and Thrive in an Era of Mass Extinction, he argues that technology is at a inflection point and that the principle technology discussion in the digital workplace at the moment is how to manage the convergence of four megatrends: cloud computing, big data, artificial intelligence (AI) and Internet of Things (IoT). While these systems are making work more ‘intelligent’ they are also increasingly difficult to manage.
Convergence in the Workplace
The manufacturing industry, for example, has been working with these trends separately for years in a number of different ways, according to Maryanne Steidinger of Webalo, and they are all starting to dovetail through the use of data. Here's how each technology is working in the enterprise:
1. Cloud Computing
Software is now being offered as a service (i.e., cloud-based, where you are essentially leasing vs. purchasing it) for the past 10 years.
The IoT has allowed manufacturers to put sensors at the "edge" of their processes, so that they have much more data from many more areas within their operations to allow them to not only more quickly respond to downtime, but also predict when those events can occur.
Related Article: 6 Technologies Needed to Power the Modern Workplace
3. Big Data
The IoT naturally generates "big data" so to deal with it, manufacturers now have options from vendors like Rockwell, Schneider Electric and others, or the analytics vendors like Tableau and Qlik that routinely deal with big data, albeit not primarily in a manufacturing or plant environment. With the analytical capabilities of big data, users can not only collect, but visualize and set up things like key performance indicators (KPIs) to monitor and respond to dips or out-of-limit processes.
4. Artificial Intelligence
For some vendors AI and machine learning (ML) go hand in hand, feeding vast amounts of data into an AI engine, and then applying context (meaning) to the data in order to understand patterns of behavior that can then lead to both good and bad events (downtime, out of spec processes).
Steidinger explained that in practical terms this means convergence happens in three ways:
1. Industrial Automation
Industrial Automation vendors such as Rockwell, Siemens, Honeywell, Schneider Electric and others are building in capabilities of cloud computing, AI, ML, big data analytics and IoT into their offerings. Some companies, like PTC and Siemens, offer "platforms" by which you can use IoT as your primary data integration and the platform handles things like infrastructure, data, integration and visualization, and it's cloud-based in order to minimize cost and infrastructure maintenance.
Related Article: Why Digital Asset Management Is Now Officially Martech
2. Purpose-Built Software
Purpose-built software companies are bringing out solutions that overlay existing architectures and integrate into existing software and hardware infrastructure, again so that there is a cooperation of sorts within the vendor community. There are AI/ML vendors, like Seeq and Element Analysis, that have established partnerships with automation vendors to make it easier for end users to buy and integrate their products.
3. Application Development Platforms
There are companies providing rapid application development platforms that integrate all of the megatrends to allow companies to develop their own task-based "apps" based on all of the data that exists within their plants/enterprise. This final example is where the megatrends truly do converge.
Convergence Is Real
Convergence is a reality, not just at the macro level, but also within specific industries and technology domains, said Doug Bordonaro, field CTO of ThoughtSpot. Technology is an enabling function, connecting ends of a value chain. Technologies between the source of data (product) and the business user who needs the information (consumer) are consolidating as the technology evolves. What used to be 10 steps with much more basic technologies is now three or four steps with much more capable platforms.
“This is the same trend Tom Siebel talks about at the macro level. Cloud, big data, AI and IoT are all converging because they're all enabling technologies,” Bordonaro said. “None of them are at the end of the value chain, so as the capabilities of these platforms improve — or become smarter — they'll naturally converge. It's bigger than this, though. You can take any modern technology category related to those four and it's likely it'll join in on the convergence. Convergence is real. It's not a new or unexpected phenomenon, but a natural outcome of the progression of technology.
In digital workplaces, all four of the megatrends he discusses are enablers of each other. IoT data couldn't realistically be stored or managed economically without the economies of scale of cloud computing or the management tools of big data platforms. And once it's stored and managed, we still need to find business value in it. Since AI is better at analyzing at scale and identifying unexpected patterns, it's a natural fit for the last few miles of this specific value chain.
Bordonaro advises enterprise managers to try to predict the future of these platforms. “Keep your eye on the ends of the value chain. Know the real reason you're investing in technology and who it's for — your value chain,” he said. “Invest in solutions that work for both of these critical factors, not just technologies in the middle that will change.”
All of these megatrend technologies have the potential to impact one of the ends of your value chain; focus on optimizing the outcome and you'll find your own way to where everything is converging.
Emerging Convergence Trends
Some of these trends are already in place and if we look at the large cloud ecosystems be it, AWS, Azure or GCP, they are all trying to be the one-stop solutions for supporting the Generate (sensor data) — Collect (storage of data) — Process (clean & transform the data) — Predict (provide inference) — Act (take action based on inference) cycle through the services they have in their portfolio, said Angshuman Patra, chief delivery officer at Ness Digital Engineering.
As predictive decision-making becomes the norm and companies try to implement AI/ML-based decisions, further shifts are happening in the industry:
1. Managing Data Sets
To build a good predictive inferences system using AI/ML, one needs large volumes of data to train, test and validate the solutions. Good data infrastructure that can be engineered quickly at scale is a necessity. Big data technologies, like Hadoop, Spark etc., that were built to handle these kinds of volumes at the scale and speed required becomes critical.
2. Cloud Compute Infrastructure
AI/ML systems processing large amounts of data needs compute infrastructure that can scale based on the computation needs. This is where cloud compute infrastructure that can be scaled horizontally and on-demand becomes important. The inherent scalability and redundancy of cloud infrastructure coupled with containerization technologies, like docker and Kubernetes, are enabling the adoption.
3. IoT Data Management
Another interesting emerging area is IoT, which is generating far more data much faster than before, and needs decision-making on that data quickly. Let’s say you have a large piece of equipment that is generating data through the hundreds of connected sensors, and the enterprise is making a decision on maintenance. You need to store, transmit, process (using AI/ML) and eventually make a decision on whether the machine needs maintenance. This process requires all the different technologies cloud, AI/ML and big data to work together to deliver the final outcome
As wireless mobile speed increases with the advent of 5G, the need to Generate — Collect — Process — Predict — Act will increase and all the different elements will start to meld with each other.