Managing Industrial Data To, From and Between the Cloud and the Edge

Modern production operations rely more on integrated data and computing power from the cloud. Yet, most legacy industrial solutions are built on siloed, proprietary, and generally inflexible systems that make retrieving data to send to the cloud for processing difficult. Adding to the complexity are new processes that require faster processing at the edge, which could mean former processes that were handled in the cloud now need to be managed closer to the devices producing the data.

Related Webinar: Creating an Architecture that Supports Smart Solutions from the Edge to the Cloud

Watch for Free

Tackling the challenges around the management of data moving from the cloud to the edge, or moving from the edge to the cloud was the main topic of the second webinar hosted by the Association for Advancing Automation (A3)’s new Intelligent Edge for Industrial Application series, which explores the Industrial Intelligent Edge, its capabilities and business impacts.

Guests for this webinar included Sam Kaira, an AI Solutions Engineer at Intel, Charlie Sheridan, Global Technical Director, Manufacturing, Automotive, and Energy at Google Cloud, and Ricky Watts, Senior Director of Federal & Industrial Solutions, Intel

The panel discussed how manufacturers can partner up with companies like Intel and Google Cloud, as well as other third-party providers, to better unlock their data from legacy machines and then incorporate that into an edge-to-cloud and cloud-to-edge infrastructure, depending on specific AI-driven optimization use cases.  Examples given of these use cases can include visual inspection AI, time-series predictive analytics, digital twin visualization and simulation, machine-level anomaly detection, predictive maintenance, and root cause identification.

“The story of manufacturing through digital transformation is a data story,” says Sheridan. “A huge amount of data is locked in different silos inside the factory, and this is where the edge comes in. A single machine could generate five gigabytes of data per week, and a typical smart factory can produce five petabytes per week. So there could be a wealth of information inside this data.”

To unlock this data, Sheridan said companies need to rely on partnerships and AI solutions to work together to enable the manufacturing workforce to easily access this data. “Everybody can’t go to college, come back with a Ph.D., and become a data scientist,” says Sheridan. “We have to make it really easy for engineers, our technicians, and Factory operations staff – we strive to turn everybody into a citizen data scientist.

One of the ways that Google has done this is through its Manufacturing Data Engine (MDE) and Manufacturing Connect platforms, which were announced in May 2022. The Manufacturing Data Engine processes, contextualizes, and stores factory data on Google Cloud’s data platform, integrating key Google Cloud products including Cloud Dataflow, PubSub, BigQuery, Cloud Storage, Looker, Vertex AI, Apigee, and others into a manufacturing-specific solution.

The Manufacturing Connect offering is a factory edge platform, co-developed with Litmus Automation, that quickly connects and streams data from nearly any manufacturing asset and industrial system to Google Cloud, based on a library of more than 250 machine protocols. Integration with the MDE unlocks rapid data intake into the cloud for processing machine and sensor data.

Insights to scale at the edge

During the webinar, Intel’s Sam Kaira spoke about how new applications that can be deployed on Intel edge servers can allow manufacturers to deploy visual inspection and anomaly detection for various production processes. Using a series of Intel-developed, open source and third-party tools, the Intel edge architecture can help manufacturing developers scale up these applications at the edge, and then connect to the cloud for further insights.

“As a form of scale we integrate Edge Insights with Google Cloud components, and then the cloud source repositories and the container registry,” says Kaira. “These are very essential for developers to be able to use these at the edge to scale, similar to how we have done in the past with the cloud. In order to scale, these components are very important and necessary.”

Additional topics covered during the webinar included discussing successful use cases of how companies were able to integrate data from legacy equipment, making sure that all of the data being moved from the cloud to the edge and back is secure, how to know which data is valuable, and what’s next for companies once they begin their digital transformation journey.

To watch the full webinar, click here to view it on demand.

BACK TO ARTIFICIAL INTELLIGENCE BLOG