distributed databases

hi,

i need some advice from people who have implemented distributed databases.

Here is the scenario:

In the diagram below is what can happen to the database being distributed at different regions under Head Quarters.

Head Quarters will need to be able to access all information from all regions.
Each region will have their own data but all table descriptions are the same across all databases except for content which will contain local data. Eg., region 1 will have customers belonging to region 1 only.

                              Head Quarters (HQ)
                         /                  \
                region1                    region2      .....  region3  ...
region 4 ...
                  /    \                        /          \
            prov1  prov2          prov3           prov4       ----> prov
means province
            /      \        / \              /   \             /    \
         site1 site2                 site1  site2    ...   ....

Questions:

1. Given each region contain only data belonging to their own region, if someone from region 2 comes to region 1 can try to do some a transaction, what is a good way to structure the distributed database ? Please also bear in mind that Head Quarters must able be able to see and access all regions.
2. There is also the issue of data synchronisation here. If the transaction was done by a user from region 2 (at a site under region1), then how does Oracle handle the data synchronisation such that the data is updated immediately to region2's database and HQ.
3. Is it possible to structure such that each province can also contain their own local data (distributing region data downwards to the other provinces) if we wish to ?

--
regards,
tony