Why not something like your mobile app pushes data to your webserver which
pushes the data to Kafka or Cassandra or any other database and have a
Spark streaming job running all the time operating on the incoming data and
pushes the calculated values back. This way, you don't have to start a
spark job for every user request. (Which will end up awful if you have
thousands of user requests per second)

Thanks
Best Regards

On Thu, Jun 18, 2015 at 4:33 PM, Ralph Bergmann <ra...@dasralph.de> wrote:

> Hi,
>
>
> I'm new to Spark and need some architecture tips :-)
>
> I need a way to connect the mobile app with the Spark backend to upload
> data to and   download data from the Spark backend.
>
> The use case is that the user do something with the app. This changes
> are uploaded to the backend. Spark calculates something. If the user
> uses the app again it download the new calculated data.
>
> My plan is that the mobile app talks with a Jersey-Tomcat server and
> this Jersey-Tomcat server loads the data into Spark and starts the jobs.
>
> But what is the best way to upload the data to Spark and to start the job?
>
> Currently Jersey, Tomcat and Spark are on the same machine.
>
> I found this spark-jobserver[1] but I'm not sure if it is the right
> choise. The mobile app uploads a JSON. Jersey converts it into POJOs to
> do something with it. And than it converts it to JSON again to load it
> into Spark witch converts it to POJOs.
>
> I thought also about Spark streaming. But this means that this streaming
> stuff runs 24/7?
>
>
>
> [1] ... https://github.com/spark-jobserver/spark-jobserver
>
> --
>
> Ralph Bergmann
>
>
> www              http://www.dasralph.de | http://www.the4thFloor.eu
> mail             ra...@dasralph.de
> skype            dasralph
>
> facebook         https://www.facebook.com/dasralph
> google+          https://plus.google.com/+RalphBergmann
> xing             https://www.xing.com/profile/Ralph_Bergmann3
> linkedin         https://www.linkedin.com/in/ralphbergmann
> gulp             https://www.gulp.de/Profil/RalphBergmann.html
> github           https://github.com/the4thfloor
>
>
> pgp key id       0x421F9B78
> pgp fingerprint  CEE3 7AE9 07BE 98DF CD5A E69C F131 4A8E 421F 9B78
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscr...@spark.apache.org
> For additional commands, e-mail: user-h...@spark.apache.org
>
>

Reply via email to