Hello, >From your described use cases I think Kafka Streams would be a good fit, in that 1) it provides higher level DSLs for windowed aggregations, and 2) it is part of the Kafka open source (coming in the 0.10.0 release, which is being voted now), and if you have your data already in Kafka it is as simple to use as a library without operational cost of another system.
Guozhang On Fri, Mar 18, 2016 at 7:11 AM, I PVP <i...@hotmail.com> wrote: > Hi everyone, > > Is KafkaStream the way to go to allow a light real time counting/analytics > that need to do things like ? : > > 1) Continuously update the total count of currently open purchase orders > for each user as purchase orders updates are passing thought the broker. > > 2) Continuously update the total count of RFXs( RFIs, RFQs,RFPs) by status > as the RFXs creation and updates are passing thought the broker. > > Any best practices or anti patterns to be aware of ? > > I am looking to KafkaStream to handle these use cases as I want to avoid > the learning curve and the operational cost of introducing another new > component like Storm and Spark. > > I am new to Kafka. > I am migrating the messaging piece of a existing application from > ActiveMQ to Kafka. > The existing application handles these requirements via scheduled jobs > (based on Quartz) that do the counting described above every X minutes. > If KafkaStream is the way to go, it would be nice to stop using the > scheduled approach. > > > Thanks > -- > IPVP > > -- -- Guozhang
