I am working on a custom peer-to-peer replication process, but can't use the real thing because not all servers are Enterprise Edition. Also there are several groups of tables that will be replicated differently, for instance some tables require replicating deletes and some don't. I would need at least two transactional replication setups and one peer-to-peer, and the amount of customization required convinced me to do it manually.
One group of tables will be peer-to-peer replicated based on identity columns, and I'm planning to use the identity seed and increment to avoid primary key violations. This would be much easier if I could use NOT FOR REPLICATION on identities and triggers, and have my sync jobs be recognized as replication processes.
So the question is how do NOT FOR REPLICATION objects know which processes are replication, and is there any way to have your own procedures take advantage of it? Can it be done without having any other replication settings in the database, so the log backups are not affected?