What is the goal? Move the (structured) data over for once and for all?
Move application functionality over?
Move ongoing changes over?
You'll have to dig into the system to figure out what it does, what it is build upon. Is there no design guide, an operations play book, backup procedures?
Most likely it is based on RMS (indexed) files. The data file would be names .IDX, .INX, or .DAT or some such, and there would be mane files, one par 'table/object'. The procedures would table about BACKUP, and CONVERT.
There would be a PowerHouse Dictionary from which metadata can be extracted with "qshow generate file" into .ph files.
You may want to look at Attunity (I work there), Connx or Easysoft to use those definitions to provide ODBC or JDBC access to the data from the outside.
Attunity has tools to bulk unload into any target DB with 'one click' once the data definitions are in place, but it is likely too costly for one-time use.
Still, if the alternative is two months of consulting/coding then a tool may be attractive.
If it is based on RDB, then you would see a few .RDB files, .RBR and .AIJ files.
There would be .SQL script morcels and operations via "RMU"
Like any other database it would include metadata and has native option for remote ODBC, or (Oracle) OCI access
Hope this helps some,
Hein.