Blog

What are AIoT “Venues”?

AIoT

What are AIoT “Venues”?

AIoT (Artificial Intelligence of Things) systems are special. They have unique characteristics that make them different than many other kinds of AI-based systems. I’m starting a series of blog posts to examine those characteristics. The name of the sequence is “The V’s” since all of the attributes begin with the letter “V”. The post examines AIoT Venues.

The multiple AIoT Venues in typical IoT systems.

Venues are all the locations of the system, and it is one of my favorite Vs. Monolithic, software systems have just one venue. Micro-service, cloud-based systems might run in multiple processes (venues), but they are all the same type of venue (cloud) and they are always interconnected.

The image shows multiple venue types. Typically, there are multiple venues (systems) at each layer, with many venues per layer at the lower layers, and just a few venues per layer on higher layers. The left-side shows a large volume of data generated by the Constrained Device Edge, with the data being abridged and compressed as it travels towards the cloud data center. The right-side shows a small volume of commands passing from the cloud down to the device edge.

The “I” in loT stands for “Internet”, but it could equally stand for “Interconnected”. Interconnected systems are necessarily distributed, multi-venue systems, and they are almost always spread across different types of venues, from on-device, to on-premise edge computer or gateway venues, to one or more venues in the telecommunications infrastructure (particularly with the advent of 5G), all the way to the cloud venue. Not every loT system needs to exploit every one of these venues, but they always involve multiple venues.

The bad news is a distributed systems is harder to design, test and debug. But the good news is it provides tremendous flexibility. “Far edge” venues (on-devices and on-premise gateways) have the best possible latency since they are right next to the environmental action. Cloud and cloud-adjacent venues have large amounts of storage and compute. The trick for multi-venue IoT systems is processing large volumes and velocities of data at the correct venue to minimize latency, bandwidth and cost. The flexibility of multi-venue IoT systems means not every piece of data needs to flow all the way from the far edge to the cloud. Superfluous data can be abridged or dropped whenever it is no longer needed. Data can be condensed or concentrated to make it high density, with the highest-value possible.

Download our description of all “The V’s” that make AIoT special.

Contact us for a deeper discussion,

    Subscribe

    Subscribe to see our posts as soon as they’re available.