Yep, we need a new in-memory engine. I can't really judge whether TC would work well for in-memory or not (we need to study up the analytics around this... aka benchmark).
My personal thought? No harm in doing it, and I suspect it will be no worse then the current engine (might even be an improvement...). Is it the final engine? I have my doubts mainly because of concurrency issues. You can certainly learn the interface this way, and... most likely this would be useful for others.
Hi!
Yep, we need a new in-memory engine. I can't really judge whether TC would work well for in-memory or not (we need to study up the analytics around this... aka benchmark).
My personal thought? No harm in doing it, and I suspect it will be no worse then the current engine (might even be an improvement...). Is it the final engine? I have my doubts mainly because of concurrency issues. You can certainly learn the interface this way, and... most likely this would be useful for others.
Cheers,
-Brian