Two-way Database Replica's - question

Discussion created by manuel.perez on May 23, 2014
Latest reply on May 24, 2014 by mousaem
I'm working on a project for a municipal client where we will be editing and updating wide swaths of their utility networks. They are currently running a 9.3 ArcSDE database on SQL Server 2005 (Old I know, in the process of going to 10.2/SQL 2012).

They are editing in a versioned environment, and since we will have multiple editors on our end, we want to edit the data the same way.

Our environment is a 10.2 SDE, with 10.2 editors on a MS SQL 2012 database.

Our environments are not connected, so I can't ask the client to create a two-way replica direct to our database. I considered using a checkout, but then I suspect if I grab the contents of that file geodatabase, copy it to SDE, edit and do our thing, I'd have to wipe out the contents of the original fGDB and replace it with what we worked on - which will look like completely deleted and then a mass of new data to the original parent SDE.

My theory is this: get the client to spin up a new 9.3 SDE thats empty, create a replica to that new database, then create a SQL .bak file to hand to me. I restore the SQL .bak file (I think 2005 is as early as we can go with a 2012 installation here), do our editing thing, then when we're done on our SDE, roll up versions, save a backup, then deliver the .bak file back to the client, overriding their replica, then they would sink the replica SDE back to the parent.
I'm sure there's a lot of assumptions here, and I'm worried about incompatibilities with the different software versions, (sde 9.3 vs 10.2, and sql 2005 to 2012. What do you guys think about the feasibility of this workflow? Am I making this too complicated and is there another way of doing this same thing?