I'd be thrilled if I could flip a switch and make our DB 10x faster. It's currently our performance bottleneck.<p>However, migrating an existing production DB (especially to an unproven technology) is no small matter.<p>And even just adjusting configuration for performance purposes can be risky. I'd prefer to do it under the auspices of a consultant DB admin with at least 10 years of experience administering and tuning this particular DBMS.<p>So in short, DB performance is enormously important, but it comes in third to reliability and stability.<p>Once your solution has been in broad production use for at least 5 years, I would absolutely consider it for a new project, though!