IIoT is developing rapidly with millions and millions of devices being deployed every day, often in critical areas of our industries. But, are our devices resilient? Can they recover quickly from connectivity problems and protect the precious data they are continuously acquiring, transforming, and transmitting?

Continuous availability of Cloud services

The answer to this question is often shifted to the almost continuous availability of our cloud services, reasoning that with a 99.99% SLA (Service Level Agreement) the service is practically always on. Until one day it is not. 99.99% SLA means 8 seconds per day of possible outage, around 52 minutes in a year.

However, two of the biggest IoT cloud providers like AWS and Azure offer a 99.9% SLA on IoT services and the hard math shows that it entails short of 9 hours per year of unavailability. I think we can agree that this must be factored in our IIoT projects. IIoT devices need to be resilient, they need a plan B, C, and even D when normal operations go wrong before declaring defeat. Some of the plans are easy to implement like “save data on the device and send it when the connection is back”, but the better ones involve a change in system architecture and the introduction of redundancy with a mix between cloud and on-premise technologies.

A resilient device

A resilient device will carefully choose what kind of data to send to different services based on their availability. It will decide to send critical alerts through the on-premise “always-on” device manager to be relayed to the local ERP while saving the machine parameters on its local storage to be sent to the cloud for in-depth analysis or KPI calculation. Whatever the plans we choose, the guiding principle should always be that disruptive technologies should always avoid disruption of service.