I want to know if my understanding is correct :
embedded : databases which are up if my application is up and are down if my application is down. I am not referring the databases used in embedded space.
in-memory : databases could be on any server; irrespective of where my application is running. Just that, these uses main memory.
Full disclosure: I represent the vendor of eXtremeDB.
Embedded databases have been around since at least the early 80's. db_VISTA, c-tree, btrieve, Empress are some of the most common from back in the day.
'Embedded database' has nothing to do with embedded systems. It simply means a database management system that is delivered to the programmer as a set of object code libraries that are to be linked with the application object code into an executable program image. In other words, the database functionality becomes part of the application itself, in the same address space. Embedded databases were used primarily for line-of-business applications in the 80s and 90s. It wasn't until the late 90s and early 2000s that embedded systems began to migrate to 32-bit architectures in sufficient numbers that database systems could be considered to be commercially viable. eXtremeDB was launched in 2001 as the first in-memory, embedded database system written explicitly for embedded systems. (8-bit and 16-bit systems do not have enough addressable memory to support a DBMS.)
An embedded database system can be either an in-memory database or persistent database (i.e. disk-based database).
An in-memory database system can be an embedded database system, or it can be a client/server database system.
A client/server database system can be an in-memory database system, or it can be a persistent database system.
As you can see, all the lines cross. You can have
And, you have have hybrids of all the above.