Lately, I've been experimenting with Microsoft R Server(MRS) and SQL Server R Services. I've come across that MRS is built on architectural components like ScaleR algorithms optimized for fast parallel execution, DistributedR parallel computing framework for managing compute resources, and ConnectR for versatile connections to data sources.
I've also learnt (and got confused) that
- ScaleR algorithms are basically functions of the RevoScaleR package built on the DistributedR framework.
- ConnectR component of the architecture is also a bunch of RevoScaleR functions / adapters to access a variety of data sources.
My question is:
What is RevoScaleR in relation to DistributedR and ConnectR?
Where does DeployR fit into all this?
In general, I want to know the high level differences between