Master repo,Work Repo and ODI installation

Question

Hi,
I have following doubts for windows flatform.
I am integrating two application systems say A(source) & B(Target) thier data located in different oracle db servers.
1) ODI server installation will be on source server/target/any other machine?
2)Where should i create master & work repositories on source/target server?
3)If my target is remote host then where can i install and run agents?
4)While creating physical schemas we can select the work schema as different,is this work schema acts like staging area? while designing the interfaces should i select this work schema as staging area? what is benefit of having this work schema different than my soruce/target schema.?
5) I read that ODI will create temparory tables while executing ODI objects each time and these tables are junk data,should i drop all these temp tables in work schema,even after or before execution of the interface?
Please clarify,
Thanks.
MNK

Answer 1

Hi,
Find my answers as below,
1) ODI server installation will be on source server/target/any other machine? Its always recommend to install ODI in your TARGET server for good performance.
2)Where should i create master & work repositories on source/target server?In TARGET server and make sure u have dedicated schema for work and master repos.
3)If my target is remote host then where can i install and run agents?Again in TARGET host , u can only install ODI run time agent in seperate servers, have a look at ODI installation guide.
4)While creating physical schemas we can select the work schema as different,is this work schema acts like staging area? Yes ODI will make use of work schema to create temparory tables ($ tables) and which will act like a staging area.
while designing the interfaces should i select this work schema as staging area? No need to select any work schema as such while designing interface, only u need to select the respective LOGICAL schema which implicitly create the $ tables in the work schema u selected in PHYSICAL schema.
what is benefit of having this work schema different than my soruce/target schema.? You will not require to have a dedicated "staging area" to consolidate your data from multiple/single source.
5) I read that ODI will create temparory tables while executing ODI objects each time and these tables are junk data,should i drop all these temp tables in work schema,even after or before execution of the interface?No need. ODI will take care of DROPPING and CREATING $ tables on FLY.
For simple data integration, 2 tables will be created at runtime, C$ and I$ which in turn will be dropped after loading to TARGET table.
Makes sense?
P.S: Experts comments are welcome.
Thanks,
Guru