To be Scheduled !!! Query Accelerator | over shared memory data retrieval

As of now even local Cross Engine data retrieval goes over loopback and network socket with extra memory copying and string to byte/byte to string conversions. There is a way to reduce the overhead using native MDB data representation coupled with over shared memory communication link.
There are number of MDB features to look at: plugin API over socket called handler_socket that was used to implement memcached client proto to access InnoDB data. Embeded MDB can do struct over presumably shm memory.