The current state of IoT suggests that any new initiative or project will likely go through a lengthy (possibly eternal-feeling) pilot or proof of concept phase. Gartner has indicated that ~80% of PoCs do not make it to production, and it is no wonder. These IoT initiatives are exceedingly complicated, even for very technically advanced teams. Most of these projects begin resembling science projects more than enterprise IT initiatives.
The start of an IoT project usually includes purchasing and sampling possible hardware that will be used to collect and process the IoT data. In this “device up” design scenario, the technical lead for the project often purchases a multitude of sample sensors and then builds a gateway from scratch, using something like a Raspberry Pi. The Pi is very flexible, easy to buy, and arrives quickly, making it appear that the project is getting off to a good start. From here, however, the project begins to slow, and the DIY process runs into common pitfalls and missteps. The Pi-based design requires hardware and software add-ons, additional programming and expertise that can be tough to find but are required if the solution is to support a stable PoC. There are also connectivity, power, support and security issues that arise.
Complicating the PoC situation further, most of these initiatives have more than 25 disparate components from different vendors involved in the complete solution all of which need to be understood, negotiated with, purchased from, installed, integrated, and maintained. The landscape is vast, each component has its own complexities and getting them to interact is a huge challenge. Finally, once your PoC is up and running you must be able to find meaningful results from the data collected.
Worse yet, getting out of PoC requires a thorough understanding of how the system will be able scale to a production level – a task that is more complicated than many teams anticipate. Based on the experience that the team has had building the initial PoC, the prospect of replicating the buildup, installation, deployment and ongoing management hundreds or thousands of times is understandably daunting. Sometimes even to the point that teams may actually prefer not getting past PoC.
In this situation, the choice of starting point may need to be re-examined. As you look toward your next IoT project it is wise to schedule more time than you think you’ll need for the research and project design phases before making any hardware purchases. Try starting from a data-down or cloud-down perspective instead of device-up, and start by clearly identifying the business result you are aiming for. When you are ready to start collecting the supporting IoT data that you need in order to gain this result it is time to start examining the hardware options available.
Here at Rigado we have gained considerable experience through almost ten years of helping IoT teams build effective, scalable solutions that get our customers into and through PoC quickly and then onto production effectively. We recommend simplifying at least a portion of these complex IoT initiatives by choosing a secure, scalable edge-as-a-service solution instead of a DIY approach. Doing so will not make your IoT initiative “easy” or “simple”, but will to get you started faster and allow you to focus on your unique applications, rather than wasting time on the ‘heavy lifting’ required to create and support your own edge infrastructure. It will also reduce your time-to-market, risk, and total costs while eliminating many of the hardware and security concerns. For more detailed information on the differences in choosing something like a Raspberry Pi versus a commercial grade IoT Gateway as your PoC hardware, check out our whitepaper “Raspberry Pi to Production”. Here’s to your next PoC making it to production!