n-tier distributed processing model

Well,

Still being a layperson myself in some aspects I'd give it a try.
In a n-tier architecture, the application simply is split into parts which run on different computers with different function. A common variety is the 3-tier architecture:
Computer A], e.g. a mainframe, runs the DATA ACCESS part of the application. It also stores the databases. So here you find the data tables and the SQL triggers / stored procedures (data-driven logic).
Computer B], e.g. a Unix box, runs the BUSINESSLOGIC part of the application. This could e.g. collect customer data from four different databases and yield a single customer 'view'. Not something on the screen, just a record which can be passed on to other tiers.
Talks to A] and C].
Computer C], e.g. a Windows client or a webserver, runs the PRESENTATION part of the application. It talks to B] only, asks for the customer view and maybe also sends modifications towards B]. The business logic tier takes care of the handling towards all databases.

Advantages