Google has deployed a pair of AI-related services to woo factories and assembly lines onto its cloud.
These offerings are: Manufacturing Connect (MC), an automation tool and data processor that supports more than 250 machine-communication protocols, and can thus receive data from a wide variety of sources; and a Manufacturing Data Engine (MDE), an analytics tool that reports on data gathered from Manufacturing Connect in what is intended to be an easy-to-use format by staff. The overall goal is to help manufacturers better understand what’s happening at their plants, and monitor their incomings and outgoings.
According to Google Cloud Tech director of manufacturing, industrial and transportation Charlie Sheridan, manufacturing businesses as a group are digitally transforming, though many of their efforts stall when scaling up.
“We believe the scalability challenges revolve around two factors: the lack of access to contextualized operational data and the skills gap to use complex data science, and AI tools on the factory floor,” Sheridan said.
Manufacturing, GCP style
Th starting point for Google Cloud’s manufacturing tools is Manufacturing Connect (MC), which Google designed with Litmus Automation. As mentioned above, it has a large library of supported machine protocols, and is designed to automatically connect to any asset that uses said protocols. MC translates the data it collects into a standardized dataset that is fed to the MDE, though Google didn’t say how the MC-digested dataset is formatted nor whether it sends all data or instead filters out what it sees as errors or invalid info.
For security-conscious organizations, this flow of information could also be a problem – especially given MC is a cloud service. Google said its cloud is designed to be run in edge environments and transmit data “via a private, secure, and low-cost connection between edge and cloud.” But it also said that its cloud supports containerized workloads and can process data directly in edge locations.
The Manufacturing Data Engine is where the analytics happen. Google said it supports multiple types of data, has its own factory-optimized data lakehouse, and contains several AI tools suited to industrial environments. Those include an integration with Google’s Looker platform, and tools to emit predictive maintenance warnings and perform anomaly detection.
Customers who’ve tested the tools include Ford Motor Company, Kyocera, and HVAC company Phononic.
What of exposed OT?
A late-2021 Microsoft study found that operational technology (OT) networks used by manufacturers are being increasingly connected to the internet, either directly or indirectly.
For this study, 56 percent of respondents said their OT network directly faced the internet for remote access, and 51 percent said theirs face the internet indirectly through their IT network. The problem, as Microsoft sees it, is that these systems aren’t secure, and leaders know it: 60 percent told Microsoft their Internet-of-Things collections and OT are among the least secured parts of their infrastructure.
With these latest Google products being cloud-based, it follows that using them requires, either directly or indirectly, connecting them to the internet and opening them up to an increased chance of attack.
Google said its connection between edge-residing MC instances and MDE deployments is secure – but it didn’t say how. It also didn’t say whether a fully on-premises version of MC and MDE would be available for businesses that can’t transmit data due to security concerns or regulations.
Pricing also only got a mention as far as the bridge between MC and MDE, which was described as “low-cost.” The Register has reached out to Google to learn more, and if it has any info to share we’ll let you know. ®