Apache Kafka introduced Mirrormaker2(MM2) with the release of Kafka 2.4. MM2 is significantly better than MM1.
I know that from the architecture standpoint, MM1 used to work using producer and consumer API whereas MM2 uses connect API. I believe that the design of MM2 was inspired from Confluent Replicator. Confluent Replicator is finely integrated with Confluent tools. But apart from that, what is the difference between MM2 and confluent Replicator?
At a high level, it's open source and the main difference is how it handles "true" active-active Kafka clusters, not just one way replication or questionable two-way offset manipulation (due to the fact that Replicator is not open source).
Thus the other difference - the support model is wider for MM2 than Confluent specific products
believe that the design of MM2 was inspired from Confluent Replicator.
False. A Cloudera/ex-LinkedIn (now Twitter) engineer developed MM2, not Confluent. Since Replicator isn't open source, it'd be hard to gather inspiration off it, other than the general idea of producing data to another cluster. Multiple open source Kafka Connect plugins already accomplish this. MM2 is architecturally different (refer KIP-382). If anything, it more closely matches Brooklin but specifically for Kafka rather than external systems
Confluent Replicator is finely integrated with Confluent tools
As of 5.4.x, so is MM2, as the Connect plugin itself is included in Kafka, and therefore any interceptors or converters that Confluent Platform also has on the classpath, would be available as well