Oracle architecture vs Sybase 11.9

Just picked up a copy of Sybase SQLServer 11 Unleashed. On pages 14-16, it goes through some of the differences between Oracle and Sybase from an architecture standpoint.

Basically, it says that Oracle is a Multi-Process Database Engine, and goes on to say, that a complete executable for the engine is fired up for every user. (Not talking multi-threaded option, just default). It has a very misleading picture on page 15, that shows several copies of Oracle processes, dbwr, lgwr, smon,pmon,arch in the diagram, implying that the entire oracle database is copied for each user. Now I KNOW this isn't true. But what truth to this is there?

Is essentially, a server process handling user requests a complete copy of the engine?
It goes on to say that as a result, Oracle hogs more memory (than Sybase), but has traditionally scaled better (than Sybase). It goes on to say that Sybase's multi-threaded engine is more efficient for a given platform, saying in effect, that it was very fast when it came out (comparatively), but is struggling to make strides in the scaling issue. Book
says that a new layer has been added to Sybase to help it with SMP, which adds some overhead, but this is not a problem for Oracle, as if there is an extra processor, Oracle will use it.

Any comments?

• Dc.