Our QE team already does a good deal of performance testing, but the tests always use a new installation of Cassandra with no preexisting data. The lack of data will certainly have an impact on overall performance for both reads and writes. We need a tool that can easily and quickly generate large amounts of data for testing. Cassandra has the CQLSSTableWriter class that can be used to directly generate data files. Building a tool on top of that class is probably the fastest and easiest way to generate test data. The tool should be easy to use in CI builds.