Big Data – Pattern detection

Objective & context

Since 2013, a major online retailing brand entrusted AKKA with their Big Data project: they wanted to be able to identify the specific behavior of its customers in real time and to interact in order to potentially trigger the sale (for example, offer a small personal discount, chat …).

Realizations

AKKA worked on the realization of an automatic detection engine of systematic characters on the flow of events and implemented a compiler capable of translating event flows into deterministic automaton.

Added value & customer benefits

The results exceeded the company’s initial expectations:

• Processing one million clients simultaneously with less than one second latency
• Processing two million with less than two seconds of latency
• Processing 5000 events per second.