Здраствуйте, господа профессионалы!
Помогите мне пожалуйста с ответом этот вопрос (MS SQL Server 7 Administration) так как в Books Online и MSDN очень мало информации на эту тему:

----------------------------------

Scenario

You have several regional offices that need to update central
headquarters throughout the day. The article of publication
that interests you consists of a table with, amongst others,
a RegionCode and LocationID fields. The replication type used
is Merge Replication.

Referring to the scenario above, which one of the following
replication MODELS is the correct MODEL?

Choice 1
A publishing subscriber model that uses a LocationID as part of the primary key

Choice 2
A distributed subscriber model that uses a RegionCode as part of the primary key

Choice 3
A central subscriber model that uses a RegionCode as part of the primary key

Choice 4
A central publisher model that uses a RegionCode as part of the primary key

Choice 5
Multiple publishers of one table using LocationID for the primary key

--------------------------------
В импортных форумах мне дали 2 ответа:

Вариант № 1: Choice 3 is suitable only for regional headquarters that are not connected to each other. Regional headquarter acts as a central Subscriber that is identified
by a RegionCode value.

Вариант № 2: Regarding to Choice 1, in real life, there is also a central corporate or global headquarter on the top of this hierarchy, that accumulates all data
from the regional headquarters, if so.... this topology helps
to minimize replication traffic between regional headquarters
and global headquarter.
This is useful when a Publisher (regional headquarters) must send data
to Subscriber (global headquarter) over a slow and expensive WAN links.

Заранее большое спасибо всем ответившим