On Thu, Oct 12, 2017 at 5:13 AM, ROS Didier <didier.ros@xxxxxx> wrote:
Hi
I would like your advice and recommendation about the following infrastructure problem :
What is the best way to optimize synchronization between an instance PostgreSQL on Windows 7 workstation and an Oracle 11gR2 database on linux RHEL ?
Here are more detailed explanations
In our company we have people who collect data in a 9.6 postgresql instance on their workstation that is disconnected from the internet.
In the evening, they connect to the Internet and synchronize the collected data to a remote 11gr2 Oracle database.
What is the best performant way to do this ( Oracle_FDW ?, flat files ?, …)
There are several ways to go about this, but for your use case I'd recommend SymmetricDS -- http://symmetricds.org (or for the commercial version: http://jumpmind.com)
SymmetricDS was originally designed to collect data from cash registers in a vastly distributed set of small databases and aggregate those results back into both regional and national data warehouses. It also pushed data the other way - when pricing was updated at corporate headquarters, the data was pushed back into the cash registers. It works with a wide variety of database technologies, scales well, and has many synchronization options. It is also being used by some organizations these days to synchronize small databases on IOS and Android devices with their parent databases back at HQ.
I first used it to implement an Oracle to PostgreSQL data migration that had to be done without down time. I've used it successfully for real time data pushes from MySQL and PG OLTP systems into an Oracle DataMart. I also used to use it for PostgreSQL bidirectional replication before other tools became easier to use. Because of its great flexibility, SymmetricDS has a ton of knobs to turn and buttons and configuration options and may take a bit to get it working optimally. If you are short on time to implement a solution, I'd suggest going with the commercial version.
