Intelligence on Things Gets Up Close and Personal

Non-stop analytics on the Internet of Things are increasingly critical to decision-making, so the intelligence performed on these Things must be close to the source – as well as residing in the cloud.

In order to leverage the fire hose of data coming from IoT you need continuous, streaming analytics. Each piece of IoT data needs to be analyzed; but if your cloud-based analytics engine should fail for any reason that data would not be included. Therefore it becomes increasingly important that certain analytics reside physically close to your important Things, as well as in the cloud.

So, along with the growth of centralized analytics in the cloud, analytics need to evolve on the “outer edge,” either on the Thing itself or on a gateway nearby. That way, Things can have their own analytics onsite to ensure back-up data collection and analysis takes place in the event of an issue.

Using a streaming analytics engine is the answer and these engines are event-driven, therefore always on; that is, always analyzing the data to detect patterns, and can be programmed to take autonomous actions, immediately driving a decision when it detects a problem.

Forrester Principal Analyst Mike Gualtieri calls the intelligence resulting from streaming analytics “perishable insights,” meaning there is only a limited amount of time before the insight is no longer relevant. So, if communications to the cloud-based streaming analytics engine fail, the Thing – or device – must have its own basic intelligence.

The April failure of the Wink Hub smart home system, designed to control everything from electrical devices to lighting, shows that communications to the cloud can and do fail.

The answer is embedded sensors with a “brain” that can store, analyze data and act on it onsite in a relatively independent fashion. So-called gateways ensure that Things are empowered with their own intelligence. This way you will never miss an opportunity, or fail to spot a potential problem, in the IoT.

Never miss a stream of data that could, for example, signal a possible equipment breakdown. Or, in the case of Wink, tell you that your house is unusable.