Building Scalable Analysis Services Applications - Microsoft SQL Server 2008 Analysis Services - page 462

<ID>Remote DB</ID>
<Name>Remote DB</Name>
<Language>1033</Language>
<Collation>Latin1_General_CI_AS</Collation>
<DataSourceImpersonationInfo>
<ImpersonationMode>ImpersonateServiceAccount</ImpersonationMode>
</DataSourceImpersonationInfo>
<MasterDataSourceID>MasterDS</MasterDataSourceID>
</Database>
</ObjectDefinition>
</Alter>
4. In the FoodMart 2008 database on the master server, create a new data source,
RemoteDS , with the following connection string to point to the Remote DB database
on the remote server:
Provider=MSOLAP.4;Data Source=localhost\Slave; Integrated Security=SSPI;
➥ Initial Catalog=Remote DB
5. On the master server, modify the Inventory Fact 1997 partition by setting the
RemoteDataSourceID property to RemoteDS .
6. Issue the Process command to the master server to process your Inventory Fact
1997 partition. After this partition is processed, you can see new the objects created
by the master server in the RemoteDB database on the remote server.
NOTE
Make sure that user account under which the master server establishes the connec-
tion to the remote server has administrative rights on the remote server. You can either
specify an appropriate user account in the DataSourceImpersonationInfo property of
the DataSource object, or make sure that the existing service account for your master
server has administrative rights on the remote server.
Processing Remote Partitions
To process a remote partition, you issue the Process command on the master server.
When the master server receives the Process command, it generates an internal XML/A
request to create the necessary objects on a remote server and sends this command to the
remote server. The request defines the type of linked dimensions used in the measure
group that contains this partition. These linked dimensions have a special refresh policy
that supports storing the dimension data on the remote server. An internal request also
defines the cube, measure group, and partition it needs to process.
In the first stage of partition processing, the remote server analyzes partition granularity
and requests all the dimension data it needs from the master server. The remote server
uses this data to decode the keys of the granularity attributes into DataID s and build
indexes. After the dimension data is loaded into the remote server, the remote server