For some applications, data making a trip to the cloud and back makes sense. The cloud is cheaper, offers more plentiful storage, and allows real-time access for people who need the data, anywhere, anytime.
Some of the newest generation of control technology is based on edge computing: computing near or at the source of data. Especially in manufacturing, it makes sense to keep data near where that data is used. Removing the latency of the cloud and putting the right data in the right hands can help prevent downtime and, if downtime occurs, can help companies quickly address the problem.
Supporting edge computing is one of today’s most notable data strategy trends: adaptive artificial intelligence (AI) systems. You’ve likely encountered adaptive AI in your everyday life using voice assistants and image recognition for face unlock in mobile phones, maybe even in semi-autonomous vehicles. In manufacturing, AI often is used in industrial Internet of Things devices in production lines, processes, and robots.
REAL-WORLD ADJUSTMENTS
Adaptive AI adjusts to real-world, real-time changes because it can revise its own code. That adjustability means organizations can become more resilient in design, more flexible in application, and more quickly able to address disruptions and downtime.
Adaptive AI-enabled control systems continuously monitor and learn the current state of operations and, with minimal staff intervention, optimally adjust the response as conditions change. Embedding adaptive AI at the device, control, and cloud layers creates three significant benefits: autonomous control, self-healing, and continuous learning.
Adaptive AI changes how fast and how well people can make decisions, which can boost productivity. That’s especially important in connected, contextual operations. Companies relying on a data strategy that involves adaptive AI will realize more flexible decision-making, as people can respond more quickly to market, production, and supply chain changes.
SIMPLIFYING DECISION-MAKING
This matters right now because decision-making is getting more complex, not less. Simplifying is not easy, but embedded edge computing can address a wide range of IT and operational technology use cases by bringing together adaptive AI, virtualization, and other IT/OT disciplines.
Adaptive AI learns patterns from experiences with humans and machines. Because adaptive AI gathers and analyzes data sequentially, it learns from new information and observations while working with older data and predictions, delivering better outcomes for those able to build, operationalize, and scale adaptive AI architectures.
ADAPTIVE AI IN THE CLOUD OR EDGE?
Adaptive AI has its uses both in the cloud and at the edge for manufacturers. Choosing which to go with in a particular situation is a function of:
- Time. Because of the data’s round trip to the cloud, the edge is best suited for time-sensitive data and when signal latency is unacceptable (or will result in losses).
- Location. Edge computing is better for remote locations where connectivity is limited or unavailable.
- Volume. The volume of data collected could be too much to send to the cloud before altering.
- Cost. You need to consider the lifetime cost of the technology.
- Storage. Will your edge devices have the space needed to store all the data your people, machines and “things” produce?
According to Gartner, more than 75% of enterprise data will be created and processed outside the data center or cloud in just the next five years. This evolution will change the way manufacturers handle data and decision-making.
As manufacturers generate high-speed industrial data at the edge, quickly discovering insights and applying those insights will become even more difficult and important. This is no longer a discussion of cloud or edge, but rather cloud and edge. The choice relies on the problem you are trying to solve and in what time frame.
Adaptive AI will be key to navigating data challenges in both realms and reaping the greatest rewards from the sources of industrial data. Adaptive AI systems already are breaking down silos and complex decision-making processes to enable new business models.
Chris Nardecchia is the Senior Vice President and Chief Digital & Information Officer of Rockwell Automation.