Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Mailing Lists -> Oracle-L -> RE: Need your idea on setting up Distributed database

RE: Need your idea on setting up Distributed database

From: Deepak Sharma <sharmakdeep_at_yahoo.com>
Date: Fri, 26 May 2000 14:20:34 -0700 (PDT)
Message-Id: <10509.106934@fatcity.com>


Very true.

My experience has been with 1 centralized (A) and 5 local databases (B). We used snapshots to achieve the distributed transaction tasks.

o For some Master type of tables (sort of static in nature such as Branch_Codes), the main table were kept on 'A' and their snapshots created on 'B'. This involved 1 dblink at 'B' to access 'A'. So, whenever there was a change in 'A', the next refresh cycle would propogate the change to 'B'. (I know your reqmt is just the other way around)

o On the other hand, for some reporting purposes, there were main tables kept on 'B' and their snapshot created on 'A'. This required horizontal partitioning of data, such that the single snapshot (there could be only 1 snapshot on 'A' for all 5 local databases 'B') would get data from all local 'B' Databases. We did this by using UNION such as:

On 'A' (This may not be the exact syntax)

create snapshot <snap_name>
select * from B1.TableXYZ_at_dblinkB1
UNION
select * from B2.TableXYZ_at_dblinkB2

etc.

This may not be good solution in your case considering the large number of local databases.

Anyway, you have some more info.

Received on Fri May 26 2000 - 16:20:34 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US