Now there are three types of stream computing framework platform
commercial grade stream computing platform IBM InfoSphere Streams and IBM StreamBase
open source stream computing framework Storm twitter with and Yahoo! S4
company to support the computational framework of business development of its own flow Baidu Dstream Taobao Galaxy stream computing platform facebook Puma
stream computing pay more attention to the timeliness of
real-time data collection tools hadoop of Flume and Chukwa
Sparksteaming micro-batch, to achieve second response, millisecond response slower than the Storm, but Storm can not be batched
SparkCore is before the Spark blog abstract data RDD is
SparkSQL is data abstraction is represented SparkSession DataFrame
SparkStreaming data abstraction is DStream stream queue based RDD
