The documentation provided by Oracle/Endeca does not provide a crystalclear definition of the MDEX, Dgraph en ENE Server. Especially de MDEX and de Dgraph definitions are confusing. One source says: "+Dgraph: An Index for every N-value+", another "+Dgraph process is a server that accepts query requests and returns results and loads the index files generated by Dgidx into memory at process startup and then listens on a specific port for HTTP requests+"
Can anyone clarify this?
If I understood the manual correctly - MDEX/Dgraph/ENE server is same. Look at your servers process and do ps -eaf|grep -i dgraph.
After oracle bought the endeca they changed MDEX name to Endeca Server.
when you do EQL it will be servered by MDEX/Dgraph/ENE servers.
Thanks for your reply. What i understand is that the Dgraph is de proces that holds the index, the MDEX is a collection of components including the Dgraph.
The term ENE Server is still confusing but i think it refers to the hardware on which the MDEX is installed and part of that installation is the Dgraph.
So they are not exactly the same but closely related.
I agree Maarten. I often use the terms MDEX and Dgraph engine/server synonymously. I'm not an Oracle or Endeca employee, so I don't know for a fact the following, but it's my best guess:
MDEX is the marketing name.
Dgraph is a historical name and the current actual name of the process that pulls the index into memory and serves query requests (on both Linux and Windows).
Ene server is a little esoteric - I'll bet it was a marketing name early on to mean Endeca Navigation Engine. The APIs have a few ENE... classes. Thus, it's a synonym for MDEX/Dgraph as well.
I wanted to add a thought to your comment " MDEX is a collection of components including the Dgraph". MDEX is the product name and package for ultimately what is the Dgraph executable.
Edited by: TimK on Sep 14, 2012 12:46 PM
It's all marketing brand "evolution" - Dgraph > ENE > MDEX. I've vaguely started saying MDEX server for the host machine and Dgraph for the actual index process, but that is subconsciously due to editing the Application Configuration (which follows that convention) so many times as opposed to any official Oracle Endeca terminology!