Edge computing is proving to be invaluable for enabling manufacturers to take advantage of modern Industrial Internet of Things (IIoT), factory of the future, and data analytics technologies. By processing data closer to the source, edge computing allows even remote facilities to overcome data transmission bottlenecks and latency issues, reduce bandwidth demands, eliminate time lags and improve cybersecurity.
That’s why nearly 30 percent of manufacturers have already deployed edge computing, and another 60 percent of decision-makers say they plan to roll out edge pilots within the next two years. But in the rush to take advantage of these benefits, it can be easy to overlook some critical factors in the deployment process that can turn an edge computing implementation into a costly nightmare. Here are a few tips and some pitfalls to avoid that will help make an edge computing implementation a success.
1. Understand the “why.” We know what edge computing can do, but what can it do for you? A successful deployment begins by understanding exactly what your goal is and what you hope to achieve. And that requires starting with a baseline assessment of where you are right now.
Before you install or deploy anything, take a detailed look at your current applications and processes. How your data is collected, processed and used today? This will help you to determine reasonable expectations to devise an edge deployment process that makes sense for what you’re trying to achieve.
2. Be realistic about resources. Legacy systems may not be compatible with modern edge computing and could require a rewrite or re-platform in order to deploy and fully take advantage of edge computing. It could mean modifying one piece of software or converting your entire facility from Windows to Linux, for example.
If you do not have the resources and staff to do that, hiring an integrator is completely acceptable – but then what? Do you have the staff to support your edge installation once it’s complete? If not, you’ll either need to hire someone or continue on a contract with a partner. Either of those is fine, but it will add to the costs of deployment and operations, so you’ll want to plan for that up front.
3. Be careful with cloud computing. While edge computing does bring data processing and computation closer to the source, you may still need to leverage some cloud resources for storage and analytics. Being aware of of how much of those resources you’re using is critical.
With on-premise computing, it’s easy to control how much you use—you have three servers and 10 CPUs, for example, because that’s how many you bought and installed. But with cloud computing’s dynamic expansion capabilities, the amount you consume can grow very quickly, unfettered, driving up the cost substantially. The cloud provider you use isn’t going to cap your consumption unless you pay them to, so you’ll either need to invest in that or conduct frequent audits yourself to make sure you’re staying within budgeted capacity.
4. Don’t bite off more than you can chew. One of the biggest mistakes companies make in deploying edge computing is going too big too fast. They simply try to do too much at once and it comes a monstrous project that takes too long, hits snags and never fully delivers on the promised dividends.
The smarter approach is a phased-in deployment. Start with a pilot to address one process or one area of the plant, prove success and expand incrementally from there. This not only allows you to apply the lessons learned from one phase to future phases but also helps to build a cadence of success with proof points along the way.
5. Edge is just one piece of the puzzle. Edge computing is not the panacea to every manufacturing computing challenge. It should be part of an overall compute ecosystem that integrates IIoT and data analytics with mobile, cloud and on-premise technologies. Edge is a critical piece of the manufacturing tech stack that can eliminate latency issues and improve productivity, but realizing that it’s just one part of the puzzle can help to set realistic expectations and ensure you get the best results and ROI out of your edge computing initiative.
Edge computing provides a powerful tool to help manufacturers leverage some of the most innovative technologies on the market to improve operational efficiency, data security and lower operating costs. But implementing an edge platform requires vast resources that range from networking and computing to storage, applications and highly specialized skills.
Having an integration partner on board who understands your goals, challenges and capacities can make for a much smoother deployment that delivers on the promise of edge computing while keeping your project on time and on budget.
Jim Demetrius is a Cloud Infrastructure Architect and Tech Guru at Tech Guidance’s parent company, TBI.