What is missing?

Oct 1, 2009 at 12:20 AM

-only works on a single machine. The techniques are designed to work cross machine and to be able to scale out automatically. That said, so far we haven't needed multi-machine scenarios and it hasn't been built yet.

 -requires a database to provide spill space for sorting. A fast, distributed file sort mechanism (a la hadoop) would be nice.

-scripting -- work is beginning on a scripting language (similar to pig latin) to enable rapid development of pipelines.

What else? If there is something Aqueduct doesn't do that you think it should let us know.