Sequences in 10g R2 RAC

Question

We are planning to implement 10gR2 RAC in the next few months.
We have 2 sites and we will be having 2 node cluster at each site and we will be using streams to replicate data across these two sites. This is 2 way/multiway streams replication.
My developers are designing an application and they are planning to use a sequence number. But they are planning to use the sequence number(at the four nodes) such a way that sequnece at node1(site A) will start with 1 increment by 4, node 2( at site A) will start with 2 and increment by 4, node 3 (at Site B) will start with 3 and increment by 4 and node 4 (at Site B) will start with 4 and increment by 4.
These sequences will be ordered and cached for better performance.
We will be using streams replication across the sites/databases.
Is this sequence mechanism good?
We will be using differrent load balancing mechanisms and in this RAC enviroment.
Please let me know if creating the sequences in this manner follow the best practices.
Thank you.

Answer 1

cache noorder is most scalable.
In our environment , we have designed as
RAC- A (2 node ): start the seuqnece with 1 and increment by 2
(1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25 ......)
RAC- B (2 node): start the sequence ith 2 and increment by 2
(2, 4, 6, 8, 10, 12, 14, 16 , 18,...,.....)