Why Industrial Asset Data Needs Edge-enabled Data Science

The New, Connected World

Increasing connectivity has delivered a wide variety of improvements to daily life from in-car assistants to the ability to instantly talk to friends and family around the world, but not without limitations and challenges along the way. Connectivity issues take many forms — from prohibitively expensive network connections to intermittent loss of connection. If cell phones still sometimes struggle to connect to WiFi, how do heavy industrial assets located at all corners of the globe maintain connectivity?

Connectivity Challenges in Industry

Take, for example, stationary mining equipment in the Arctic Circle. These mining assets produce high-frequency vibrational data filled with machine health and efficiency information. The enormity of the data coupled with the cost of internet transmission in an area with limited access makes it near impossible to realize the value of that data through traditional cloud-based data science architectures.

Different challenges come with mobile assets, like locomotives transporting precious metals in the outback of Australia. These locomotives travel in and out of network coverage, and though it’s possible to intermittently deliver data to and receive insights from the cloud, it’s very expensive. In addition, many locomotives carry such precious cargo that the cost of machine failure is simply too high for them to only be monitored intermittently.

The Value of Edge-Deployed Data Science

Businesses with remote assets need to be alerted in real-time, before they return to network coverage, to avoid catastrophic and costly failures. Edge-deployed data science augments existing cloud-infrastructure, increasing the speed of analytics and reducing load on cloud networks. The basic idea of data science on the edge is to move the computations that create insights to the data source. Doing so enables businesses with limited or no network connection to gain real-time data insight and take immediate action.

Improved Cybersecurity

Data science at the edge is also more secure, which helps companies with highly sensitive data gain insight without exposing themselves to cybersecurity threats. Cybersecurity is an undeniable liability that every company must keep top of mind. By moving from a cloud-based to an edge-based analytics implementation, less data leaves the premises, making operations less vulnerable to cyber attacks. Edge-deployed data science offers the rewards of advanced analytics while mitigating cybersecurity risk.

The Future of Connectivity

Edge-deployed data science is definitely in the early stages of development, and deployment remains somewhat customized. Limitations include customers with disparate data needs, network access and security concerns. This means that most businesses require solutions that fall somewhere along the cloud-edge spectrum, with the most optimal solutions being hybrid. This is a truly remarkable feat of engineering that requires careful collaboration with industry experts and device manufacturers. And data scientists like myself are adapting existing algorithms and even inventing new ones suited for edge computing. In the not-too-distant future, we expect a level of maturity similar to cloud computing, in which all of the estimated 44 zettabytes of digital data can be analyzed in real-time providing efficiency, reliability and security.

Learn more about how Uptake's products turn edge data into actionable insights.

Michael Cantrell is a Data Scientist at Uptake and leads the data science team's Edge efforts.