twmarshall / tdb Goto Github PK

We are currently using an older version of sbt due to a bug in the latest release that makes it incompatible with REEF. We should remember to check for the bug being fixed in a few months and move up to the newest version of sbt then.

We would like support a Hive-like SQL API on top of incremental change propagation.

We should have an example Mutator that takes an adjustable algorithm as input and then interactively allows the user to modify input and have the results printed out. Given that Scala has an interpreter, I think we can just leverage that.

We need benchmarks for comparison, both written in Scala and implemented on other systems (DBToaster, Spark, etc), implementing algorithms which we have also written incrementally.

We need some way to bulk import data from a file on the client machine (and maybe eventually from a file on HDFS).

Right now, the key-value store is just an in memory hash table controlled by a single actor. We need to make it distributed and persistent, either by swapping in an existing system such as WireTiger or by building something of our own.