Evaluate Weigh the pros and cons of technologies, products and projects you are considering.

Pros and cons of connecting industrial sites to the cloud

Scott Berg is the chief executive officer of ServiceMax, an operating unit of GE Digital. In his daily work, Scott and his team receive questions from customers around implementing industrial and manufacturing IoT systems. This column serves to share some of those questions from the field and provide actionable advice for readers.

Dear Scott: What are the pros and cons of connecting industrial or manufacturing sites to the cloud? I’m trying to understand the different considerations for a public cloud versus a private/on-premises cloud. Is it really all about security?

Dear Reader: A decision between public and private cloud deployments is much more than a question about security. Every business should care about the security of their information — it’s a shared priority for all of us and is table stakes in this day and age.

I often counsel customers around cloud considerations including data volume, proximity and immediacy of availability. When choosing a cloud framework, your decision really boils down to the unique priorities for your operations at hand. Considerations such as how much data comes from a given machine and how fast you want your systems to respond to that data are key. Security and personal access to that data are important, but not the first point you should necessarily consider.

When thinking about the equipment in your facility, ask yourself the following questions:

  • Is there a need for the data servers that inform your machines to have physical proximity?
  • Is there an immediate benefit to the speed and immediacy of real-time data transfer between controlling databases and the machinery itself?
  • Is there a constant stream of real-time data updates or two-way communication occurring between a given piece of equipment’s sensor array and your company’s core informational databases?

Security must come into it somehow though, right?

The debate between public and private cloud rollouts is something a lot of CIOs, IT and operations leaders grapple with. People tend to oversimplify the pros and cons and look at this as a straightforward data security question, as if the private cloud is simply more secure than the public. Security is absolutely a consideration, in terms of who can gain access to data, but what many people don’t realize is that there are ways of securing a public cloud just as well as a private cloud. Data security not only entails controlling who can access data, but also making sure it is backed up and won’t be lost. If you store data in a public cloud, you may lose complete data access security, but you gain data backup security. For example, if you lose a USB stick or a laptop, that locally stored data is still saved in the public cloud.

When is public cloud better than private cloud (or vice versa)?

Let’s think through two hypothetical scenarios. If you are the IT director for a medical device company, imagine you have a fleet of 1,000 MRI machines in operation across the U.S. For your purposes, a public cloud system might be perfect. Your goals here are to maximize uptime and understand usage patterns so you can improve MRI reliability for hospitals over a period of years. The advantage of a public cloud is being able to get data out of each physical machine location and have that information live centrally, so you can access that data across a variety of locations or even worldwide.

Now, contrast that MRI scenario with an additive manufacturing context, where the data being collected is critical in real time. In this scenario, an industrial edge computing platform may be the way to go, because you can put ingestion, storage, analytics and even applications right next to the manufacturing machine. In this situation, an edge system gives you the advantages of speed and the ability to react quickly in high data volume scenarios.

Now play that forward and imagine you have a huge fleet of 3D printers producing highly specialized parts for a customer with incredibly stringent data management requirements across a complex, distributed supply chain. This is where a private cloud system might make the most sense. You may connect and monitor edge devices, run application performance management programs and execute powerful analytics against large data sets in your own private cloud — all wholly separated from external threats.

Over the years, I have learned that users need a selection of systems for various points along the industrial compute spectrum: from the edge to the cloud, and some hybrids like internal, private clouds. That’s why my company developed offerings to give businesses flexibility when and where they operate and make it easier for customers requiring an on-premises industrial IoT platform.

At the end of the day, you know your business better than anyone — that makes you the best person to judge whether a public or private cloud system is best. There are pros and cons to both, but what they look like will differ from one business to the next, and from my business to yours. Remember to look through multiple lenses — data volume, proximity, immediacy and security — and allow your practical needs to steer you towards your answer.

Do you have an industrial IoT question for Scott? Email us and your question may be answered in a future post!

All IoT Agenda network contributors are responsible for the content and accuracy of their posts. Opinions are of the writers and do not necessarily convey the thoughts of IoT Agenda.

Data Center
Data Management