i'm not thinking, it's a open MDEV that i read and was thinking about it (i don't know why i was thinking about it kkkk, no work and free time =] )

but i was reading about memcached some months ago and there're some memcached daemons that runs on memory and disk, and since query cache is something like it (a cache), why not use the storage engine of mysql to put query cache? that's the point that i was asking my self

about the bottleneck, i think it's something like a table lock, the query cache is locked and no one more can write to it, when two or more queries start writing to query cache the bottleneck is write to query cache, but i don't know internally hwo it's done, and i never see this problem in running servers, i don't know why should partition the query cache, maybe just for reduce locks?

i know the query cache flow via debugger, i don't have an idea about how everything works in this scenario, there's some prunes, but i know what docs say and a little about i read at sql_cache.cc