IoT intelligence can be developed at three levels. If we measure the distance from centralized equipment, the first level is the farthest and is defined by the ability of the devices to process information, the second level give the gateways (gateways) grouping traffic from different devices and the third level is in infrastructure businesses or centralized platforms to devices and gateways send information.
The first two can say they are the Edge and the third is the Cloud, a system that has only Cloud is a centralized intelligence system, while other processes have on the Edge Cloud is called distributed intelligence.
Today most of the analysis of IoT happens in the cloud, this is due to the supply of suppliers / manufacturers are based on it, that its centrality provides more computing power and storage and unique development.
But increasingly it makes more sense to think of what kind of analysis can be done in the "Edge" (perimeter).
In the analysis we have 4 phases: descriptive, predictive and prescriptive streaming (as you can see in more detail: data analysis IoT ).
The discussion is now know which of these phases can be made in the "Edge" (perimeter) and what can be done in the Cloud (centralized).
Value chains Analysis on IoT's consist of the following five terms:
Where Analysis on the Edge can provide value ?.
So we can define three types of Edge: Far (Extreme Edge), close (Near Edge) and fog (fog).
Extreme Edge .- A sensor or a machine working separately, today these devices have limited their data collection capabilities, security and memory, perhaps in the future with the advancement of semiconductors may have more capabilities and certain functions of the analysis could be made there.
Near Edge .- Refers to the gateways connected to devices where it is most of the analysis of the Edge because it is the only place where suppliers offer.
Fog .- Here refers to telecommunications infrastructure where analytical skills that are very defined according to local and regulatory requirements, and having a clear example of application to smart cities (Smart Cities) is allowed.
In the three types of Edge you have to work to build value giving more capacity for analysis, so that it can be more autonomous, fast and safe, and go to the Cloud with really valuable data and when necessary.