On Mon, Oct 31, 2011 at 6:50 PM, John Lenz wrote:

4) My code processes hundreds of messages. Every message is processed in

...

exactly the same way as the others. So the code can be easily parallelized. Any Haskell frameworks that will allow me to run this code in a simple concurrency model?

Yes, there are many options.

http://www.haskell.org/**haskellwiki/GHC/Concurrencyhttp://www.haskell.org/haskellwiki/GHC/Concurrency

John, thanks for detailed reply! I am looking at Haskell Concurrency wiki, but can not figure out which framework - STM, sparks, threads, etc. Amazon AWS will be able to scale? As far as I know, to scale CPU and program memory in Amazon, all you can ask from AWS is to start some number of additional extra VMs. Every VM contains a complete image of your OS and executables, all images are exactly the same. That's fine with me as currently all my workflow tasks are performed by separately compiled Haskell executables communicating via regular files. So to reformulate my question: - Does any Haskell framework exist that allow to orchestrate separate processes (NOT threads that share the same process memory)? On the other hand, in case there is a way to make Amazon AWS to scale Haskell STM, sparks, threads, etc. - I would happily rewrite my Haskell code to use these frameworks. So my second question: - Is there a way to make Amazon AWS to scale Haskell STM, sparks, threads or any other Haskell concurrency frameworks?