For now SparkStreaming supports three main business scenarios:
➢ stateless operation : only concerned with real-time data in the current DStream . For example: Only do check the correctness of the data in the current DStream. (Nothing to do with the next piece of data) suitable for warning system
➢ stateful operations : When DStream stateful operation, the data before the need to rely on . For example: site statistics of each module of the total number of visits . Cumulative multiply tired Summary
➢ window operation : operating on the data within a specified time period DStream range. For example: the need to count the number of each module to access the website in one day. hot spot
Acquaintance flow calculation
View this example, convection calculation intuitive understanding
◆ Start Service
◆ nc -lk 9999
◆ RUN-Example streaming.NetworkWordCount localhost 9999
nc netcat command network tools