We have used Streams. We had a lot of network latency in Streams. All the applications went to one pipeline. With GoldenGate, we can create multiple pipelines.

We moved to GoldenGate because there is less downtime for the end user. Everything is replicated in the new environment. It also provides consistency. 

Before switching to GoldenGate, we had some experience with the AWS Database Migration Service. 

Previously, I used to use Oracle Data Integrator and a solution from IBM, but I don't remember its name.

My company uses Oracle GoldenGate because it is bundled with other Oracle products.

I was not at the company when the organization chose this solution. I'm not sure what was used previously. 

We have used many other tools in parallel with Oracle GoldenGate, such as Qlik Attunity for the replication.

Quest Shareplex, because of licensing costs.

This is a new solution for us. We did a proof of concept for all of the systems out there and discovered that, from a performance perspective, Golden Gate certainly met all of our requirements, which is not necessarily the case with some of the rivals out there.

Oracle has implemented a slightly different technology. Instead of scraping the redo logs, the way the traditional replication of this type of system works, it's actually using a new technology, with its integrated Extracts and Replicats, to effectively use Oracle's streams to double the performance. It can literally run twice as fast as anybody else. We needed that to keep up with our transaction volumes.

For me, the most important criteria when selecting a vendor is that it does the job.

At the scale where we are now, we have worked with Attunity. It is a similar CDC software data extraction. It's more straightforward and simple to configure, however, when you scale with the 3000 data sources, the system starts being unmanageable. And the total cost of ownership is higher than OGG.

I've only ever worked with this solution. I haven't tried any similar solutions in the past.

Prior to this solution, we had a package that used an SQL process. It had come from Oracle Scheduler. We switched because it negatively impacted our production database, making it very slow. If there was downtime and the production database had transactions then it would take days to recover from the lag.