Major refactor of code to make it work in Spark Cluster mode

[pravinbhat]

What this PR does: Major refactor of code to make it work in Spark Cluster mode

• Makes CDM work when deployed on a Spark Cluster (does not report metrics counts correctly in this mode)
• Helps make CDM more performant for large migration efforts (multi-terabytes, several billions of rows) using Spark Cluster (instead of individual VMs)
• No feature changes (other than removing unused Guardrail feature within Migrate and DiffData)
• Fully backward compatible (other than the changes to Guardrail feature mentioned above)

Note: This PR/release will have a bug (only in cluster mode) that it does not report metrics counts when deployed on Spark Cluster, however it does the underlying tasks (Migration, DiffData, GuardrailCheck) as expected. When deployed as a Spark job on a single VM (not cluster mode), everything (including the metrics) will work as expected. So no impact to existing customer who use the tool in the single VM fashion

Checklist:

• Automated Tests added/updated
• Documentation added/updated
• CLA Signed: DataStax CLA