While the data generated from an Internet of Things (IoT) deployment holds a lot of promise, it can also hold a lot of complexity and risk. At the moment, 58 percent of IoT projects fail and only about 25 percent of companies derive business value from IoT. To improve success rates, organizations need to carefully think through a number of elements before writing a Request for Proposal (RFP). In no particular order, here are five of the ten factors to consider for IoT at the edge, taken from my paper, “The Top 10 Considerations for IoT at the Edge.”
Light edge vs. heavy edge
As you go about planning your IoT implementation, you have flexibility in where you perform analytics, machine learning, etc. However, performing these functions at the edge requires additional infrastructure (read: investment). A light (weight) edge is a relatively simple solution that gives a user compute, storage, security, and network functions to allow them to connect, run, and lightly process data or send outcomes to be handled by the cloud. Compare this to a heavy (weight) edge solution that has a full stack of technology that enables “cycles” – heavy applications such as analytics, machine learning, etc. – to do actual work at the edge of the network. If you prefer a thinner edge, you can collect the data and get it out as quickly as possible or perform preprocessing at the edge with minimum compute and send the rest to the cloud. The choice is yours, but regardless of what you select, it’s important to understand the implications of that choice.
Data management and analytics
Much like light edge or heavy edge, the decisions you make about data management and analytics will have far-reaching implications. There’s a lot to consider, including how much data you have, what you want to do with it, whether you’ll run artificial intelligence or machine learning to perform analytics and predictions — and, of course, how you’ll manage it. You need the technology, but do you also have the people and are they properly skilled?
This is a no-brainer, right? An IoT deployment requires enterprise-grade security controls and best practices. Think network segmentation and defense-in-depth. At a minimum you need security on the hardware, network, applications, and data layer. Don’t cut corners here!
Data governance, like security, is essential. Governance is about protecting your intellectual property — IoT data — by ensuring that the right data is going to the right place at the right time and is accessed by the right person and the right IoT asset. For example, you get to choose if a piece of data goes back to a robot vendor or a cloud service provider. Not only do you have the right to make these choices, but you must do so. If you don’t, you risk the wrath of regulatory compliance auditors, privacy laws such as GDPR, or even industrial espionage and sabotage.
Integrating the technology that comprises an IoT solution at the edge is no easy feat. You’ll likely have hardware, software, a gateway management platform, compute, storage, application, and data management solutions — all from different vendors. You’ll be lucky to find a systems integrator who can do this (pricey) work. To reduce complexity and costs, look for a vendor that provides a simple solution that does the integration for you.
As I mentioned at the opening, these are just five requirements taken from a recent guide I put together. If you’d like to dive deeper into the five requirements I’ve introduced here and five additional that will help reduce the risk of your IoT deployment, read “The Top 10 Considerations for IoT at the Edge.”