Big data processing method, device and system

The invention relates to the field of data analysis technology, in particular to a big data processing method, device and system. According to the method, a corresponding analysis terminal with the least user equipment number is determined according to the user equipment number corresponding to each analysis terminal, and original data is sent to the determined analysis terminal to trigger operation by the determined analysis terminal, wherein the operation comprises the steps that data conversion processing is performed on the received original data to obtain corresponding set state data, and the set state data is sent to a Kafka cluster so that a user can acquire subscribed set state data from the Kafka cluster. By the adoption of the method, dynamic load balance between all the analysis terminals is realized; and furthermore, the user can acquire the set state data corresponding to all user equipment in real time and detect whether user equipment with a running anomaly exists among all the user equipment i