What to Look For in an Edge Computing Solution
A good edge computing solution should be able to both show straightforward, immediate benefits while also enabling organizational growth in the long term. It should output useful data and reports to all existing systems. Perhaps most important, it should be optimized for the requirements of the specific production environment, providing a useful tool for line engineers to improve overall operational efficiency (OOE).
An edge computing solution will be a long-term partner, so this is a significant acquisition decision. The wrong choice can significantly constrain future growth. Fortunately, there are some clear criteria for judging and comparing edge computing solutions for the production line and the plant.
Immediate functional improvement
The solution should show immediate benefits to the engineer. It should not require any changes to either equipment or procedure. Edge computing should take the data smart cameras and barcode readers already generate and integrate it, filter it, and turn it into a form useful for making decisions about the production line.
This should be easy and straightforward, fitting into existing workflows without the need to acquire additional skills, detailed knowledge of vision systems, or the ability to interpret large amounts of new data. It should not make the controls engineer’s job harder or more complicated, and should not require attention already budgeted for existing tasks.
The edge computing device should plug into an Ethernet switch, get connected to a programmable logic controller (PLC), and be useful right from activation. It should be something for the engineer to use and experiment with, without negatively affecting existing metrics.
Scalability is not just a buzzword. Many perfectly workable small-scale solutions can run into serious trouble as the number of users and resources, the physical distance over which they work, or the admirative overhead grows. Such trouble can create additional costs and delays, or even tank the whole project.
The ability to start small is just that—the start. As the production line using an edge computing platform starts to report on its improved productivity, there will inevitably be demand to expand and implement edge computing more broadly. With each extension of scale, there is a wider range of data, more interconnections with other systems, and a greater risk of data overload. The chosen solution will need to handle that kind of growth without requiring custom modifications, costly additional systems, or third-party consultants.
Data output flexibility
The data generated by an edge computing system can be used by many different software and business systems in a modern plant. Uses for this rich data will grow and evolve over time. An inflexible or closed architecture that cannot export data consistent with a wide range of protocols will quickly end up imposing delays and costs, adding demands on overburdened IT departments. The edge computing system should work and play well with all other existing systems in the plant and company.
Robust reporting capabilities
Production engineers often have to pull spreadsheets, clean the data, and generate pivot tables and other data displays just to organize and understand information about their own line. This is time consuming and often excludes valuable information.
An edge computing platform should be able to automatically generate regular reports on KPIs, with no additional work required. When higher-level decisions are required, it should be fast and easy to extract crucial information and send it on to management in a format convenient for their use.
Such a reporting capability allows an engineer to keep a pulse on a significant number of lines. If there is variation between one line and another, the variation can be reviewed to determine the underlying causes. Alerts can be set to notify if performance measures, such as read rate or throughput, drops below certain thresholds. Flexible reporting gives the engineer a finger on the line’s pulse.
Built for your specific application
Many available edge computing software solutions are generic, developed entirely by IT engineers. They are billed as applying to almost any domain, as a top-down solution that does not need to consider the day-to-day details of work on the production line.
However, they can only ignore mundane day-to-day details by requiring the user to customize the solution to their own circumstances. This often requires a lot of knowledge and attention to modify vision systems, statistics, and user interfaces. The third-party systems integrator who inevitably gets called in will mean additional expense and delay, and will likely still not provide an optimal solution.
This gets things completely backward. Your edge computing implementation should show an understanding of production line operations and their specific requirements, while handling user interfaces, displays, and reporting to minimize the amount of extra work or frustration imposed on the engineer.
Fortunately, such a production-specific edge computing finally exists.
Know what you need
Such a system can really change the way you work for the better. Consider how you would like to improve your understanding of your line’s operations, and what that would allow you to do. If you’re the first one to try it out, you can be the one whose OOE and error rates are the ones other lines are trying to beat. To learn about the Cognex Edge Intelligence solution, download the datasheet.