In the 1940s when my dad was a kid, a tour of the factory floor of almost any manufacturing company would have been filled with people. Lots and lots of people. In an automotive plant, he would have seen people using welders, assembling frames or body panels, using impact drivers to attach leaf springs, running presses and casting machines to make solid steel axles. People would be building cars.
And those people were each responsible for their part. They spoke up when their welder wasn’t working correctly or if whatever tool they were using was worn, broken or not quite right. And there were lots of other people too — people responsible for machine calibration and maintenance; people who kept the line assembly (literally) moving; people who replaced parts on machines, made sure the machine was working and logged reports on the failures and fixes.
Then still more people collected, coordinated and aggregated those maintenance reports (on paper!) to set machine inspection schedules, evaluate potential changes in assembly processes and recommend a host of other potential improvements. (Those reports were probably bound into books and that data summarized and analyzed … eventually.)
Wind the clock ahead to today. If I took my boys on a tour of a modern automobile manufacturing plant, what might they see?
They would probably see robots, lots and lots of robots — robots that weld, that attach fasteners, that move fabricated parts into the right places in the assembly line, and that check and sort incoming inventory, putting it into stocking bins that other robots use. My boys might also see some pretty cool behind-the-scenes stuff like really smart machines that monitor other really smart machines. They might see machines that diagnose impending failures and report them (along with diagnostic details); that collect, review and coordinate failure data to find trends; and that convert the trend data into dashboards that are simple for humans to understand.
They’d see some of the best that technology has to offer. Would they still see people?
Manufacturing has changed to be smarter
The short answer is yes, they would. The longer answer is that they might not see people holding welders or running casting machines. They may not see people using their own muscles to move inventory.
Instead, they might see people installing and programming large, complex machines. Rather than huge bound volumes of factory diagnostics and repair logs, they would probably see state-of-the-art storage platforms that support lightning-quick data analytics and customer preference trend reporting (tied back to automated ordering, payment, receiving and stocking support systems) on their tour of the site’s data center.
Remember the people who used to replace parts on the machines, make sure the machine was working and log reports on the failures and fixes?
Those people would still be present. But instead of service teams having to wait for a machine to falter when a part needed to be replaced, these people could use the massive data collected in the smart factory — including sensor data that could listen to the machines — to predict when the machine needed service. And the best part? By servicing the machine before its failure, they could prevent catastrophic downtime in the factory.
My sons might see smart manufacturing in action — with highly trained, highly skilled people using, managing and supporting some of the best that technology has to offer.
These changes also uplift human skills
Changes in people and the skills they use every day in a smart factory aren’t the only difference they’d see. Smart manufacturing can waterfall into broader areas like increasing people’s skills in field service and “back-end” support.
A traditional field diagnostics and repair visit
I recently needed on-site service for my home satellite TV system. (I don’t remember the exact issue, but the phone support team was unable to diagnose the cause.) So I scheduled this service.
As sometimes happens with field service calls, the scheduling team gives you their best estimate for when the service technician will arrive (and they give the service technician their best estimate of the nature of the problem and time needed to address it). Despite these best estimates, the customer’s experience may be negative — arrival and duration times are uncertain, the necessary diagnostic equipment isn’t readily at hand, or on rare occasions, the issue escalates and additional services are required. The result? A delayed remedy or a second visit.
In the case of my system, this traditional visit led to feelings of frustration and inconvenience, and not only for me. I could see a similar expression on the service technicians’ faces; they didn’t want to come back either. But they did — and this time, the problem was fixed.
A mobile service/repair workforce enabled by smart manufacturing and data access
Now imagine that same service call, but this time the field team is connected to a smart manufacturing floor using high-speed wireless (maybe 5G). Imagine that, in addition to their personal skills, experience and knowledge, the technicians had access to all the diagnostic information and issues resolution data — in real time, fully integrated with the current and historical data from the manufacturing floor.
Image they were also equipped with a recommendation engine that sorted all the potential known paths to address a problem based on historical diagnosis and fix effectiveness — with direct tie-in to the data coming from the manufacturing floor. Field service technicians could input symptoms, tie those to manufacturing data and see the most popular, most successful remedies in real time. All these remedies would have been developed from the accumulated knowledge of all the service calls to date via a flash-based NoSQL-powered back-end platform. Technicians would get the latest fix recommendations, correlated with monitored factory and machine data to further enhance their human diagnostic capabilities. The customer would likely be satisfied by the result without necessarily realizing all the technology being used for the fix.
The newest, most successful fixes, patches and advisories would “bubble up” to the top of the recommendations, meaning that the results of my service call would directly enhance the next customer’s service call. Now imagine easily scaling that technological force to a business level.
Smart manufacturing builds better with Micron mainstream 7300 NVMe SSDs
In our smart manufacturing example, we have a smart manufacturing line (with machine monitoring, corrections and updates) feeding a high-performance NoSQL database back-end in near real time.
Inside, we’d have a factory whose monitoring data, orders and system status are available in real time. Our manufacturing efficiency could climb because of our understanding of machine calibration, variance and trends, order entry and status. Imagine a supply chain that cross-ties incoming orders with logistics to manage the smart factory, which is in turn tied to an active, up-to-date knowledge base used by field service teams for a stellar customer experience.
In the field, we’d have support teams equipped with the latest, most successful resolutions that are all tied together and communicated via 5G wireless, enhancing service calls and first-time resolution.
What does a NoSQL performance look like?
To better understand how our 7300 NVMe SSD processes common NoSQL database workloads, we used a standard benchmark (details on Yahoo Cloud Serving Benchmark and workloads are here: https://github.com/brianfrankcooper/YCSB/wiki/Core-Workloads) to measure the advantages the 7300 can bring that an enterprise SATA SSD cannot.
This micron study (available on micron.com) shows how the 7300 can boost NoSQL database results.
How about we wind the clock ahead again — to a time when my grandkids are taking that same tour of a modern manufacturing plant. What might they see? My hunch is they will still see people doing the advanced jobs needed to keep a complex factory running. My hunch is that people are here to stay.
