This page describes the steps planned for getting to the first release of the Firewater project. They may change.
Milestone 1: First Development Checkpoint
Criteria: in a state where someone other than JVS should be able to sync the code, build, and run it in a development (not packaged) environment with multiple nodes.
- minimalist home page and sf.net project page
- create email@example.com mailing list
- DDL support for creating storage nodes and partitions
- distributed query support for everything except joins
- build doc
Loading will still be "manual" at this point (inserting data directly into partitions by executing SQL against each storage node individually).
Milestone 2: Benchmark Readiness
Criteria: able to manually load Star Schema Benchmark (SSB) data and run queries
- DDL support for choosing replication vs partitioning on a per-table basis
- distributed query support for joins
- benchmark setup and execution doc and scripts
--Jvs 22:03, 29 August 2010 (EDT): we can has
Milestone 3: Kettle Plugin
Criteria: able to run a Kettle job to load data after registering only the distributor node (not the storage nodes)
- remote bulk loader support for storage node target
- plugin source
- plugin build and usage docs
Milestone 4: Packaged Prerelease
Criteria: binary packaging has been developed to the point where someone other than a developer should be able to download the system, install and test it, and provide feedback.
- beefed up home page, project FAQ; update LucidDB home page and FAQ to explain the relationship between the two projects
- getting started doc
- publish download link for binaries
Milestone 5: AMI Prerelease
Criteria: AMI packaging has been developed to the point where someone other than a developer should be able to download the system, deploy on EC2, test it, and provide feedback.
- AMI deployment doc
- publish download link for demo AMI's
Milestone 6: Release
- release source and binaries on Sourceforge
- publish demo AMI's
- documentation for what works and what doesn't