Distributed Analytics
Smart Agora can provide live collective measurements from users who are localized in points of interests.
In other words, the citizens receive real-time aggregate information, e.g. average, sum, min, max, etc., of the choices that other citizens make in points of interest. The collective measurements update while citizens join or leave the points of interests, therefore, real-time aggregation measurements is a result of citizens’ mobility and localization in points of interest.
The update of the collective measurements can be a result of complex localization logic referred to as collective measurements maps. A one-to-many measurements map updates collective measurements each time a citizen is localized in one of the points of interest (logical disjunction OR).
In contrast, a one-to-one measurements map updates collectives measurements for each localization in a certain point of interest. Therefore, each point of interest has its own collective measurements. More complex collective measurements maps can be designed depending on the application scenario, for instance, two map layers, one for bus and another for tram stations.
Smart Agora can perform such real-time measurements in a fully decentralized fashion without relying on a central trusted third-party that collects all data to perform the computations.
Measurements are crowd-sourced in citizens’ networked devices and are performed in a peer-to-peer fashion.
These decentralized measurements are especially challenging to perform when devices join and leave the network meaning when updates of the collective measurements are subject of witness presence, i.e. localization in points of interest. To tackle such challenging computational problem, Smart Agora integrates with DIAS, the Dynamic Intelligent Aggregation Service that is capable of performing self-corrective operations in collective measurements when nodes join, leave or fail in the network. This is possible by using a distributed memory system based on probabilistic data structures, the Bloom filters. More information about DIAS is available here.
