Abstract:

According to its timeliness,big data processing systems can be categorized into two groups,namely batching big data processing and streaming big data processing.Both systems mentioned above are unable to meet the real-time requirement for censoring and query analysis tasks.To this end,the “stream cube” real-time data analysis technology and platform were presented,which can perform timely query with low lag.Currently,this technology has been applied to many fields,including financial risk management,anti-fraud as well as web bots defense,and offers promising prospects for further applications.

Key words: streaming big data, streaming processing, incremental computation, time series processing