Many core systems are unlikely to be migrated to an Azure SQL database anytime soon, particularly large applications and being able to use Fabric mirroring to make that critically important and valuable data available in Fabric with little effort would enable the rapid integration of this data into analytics. Removing the data movement friction and letting us focus on modeling and analysis right away would provide substantial benefits.
We are actively working on bringing the Mirroring for SQL Server on-prem to preview. It is currently in our roadmap.