What to do with the Falcon engine?
After Oracle/Sun, Falcon’s ‘InnoDB replacement’ story faded — keep it as a pluggable niche engine, not a discarded corpse.
Falcon existed largely as a strategic non-Oracle InnoDB alternative. With Oracle acquiring Sun, that political reason-to-exist looked spent — and the engine was already late.
Engineering case for keeping it
Storage engines earn niches from architecture, not slogans. Falcon’s value (if any) was specific behaviour under specific workloads. Dumping it wholesale wasted that work.
Practical positioning at the time:
- Ship Falcon as a pluggable for the 5.1 line people actually ran (not only 6.0 trees)
- Separate VCS project so outsiders could test once crashers were under control
- Treat stability and public testing as the gate — not roadmap theatre
Whether Oracle continued Falcon is history. The reusable lesson: pluggable engines need distribution channels that invite testing, or they die of neglect regardless of design quality.