Enhance SqlGeopartition app to segregate reads/writes to the local "region" only

[ajcaldera1]

Now that we have leader_preference in tablespace groups, and there are more geo-partitioned use-cases coming up where customers are trying to read/write only to a specific region partition, we should enhance the SqlGeopartitioned workload to do reads/writes only to the connected "region" instead of randomly picking a region based on the modulus of the number of overall partitions.

E.g.
Consider the case where you have highly distributed regions:

ap-northeast-1
eu-west-1
us-east-1

In a geo-partitioned workload, where "fast" writes are required, you could set up a table that is comprised of multiple partitions that belong to 1 region only.
E.g.

ap-northeast-1a, ap-northeast-1c, ap-northeast-1d
eu-west-1a, eu-west-1b, eu-west-1c
us-east-1a, us-east-1b, us-east-1c

As it happens right now in SqlGeopartitionedWorkload, these would be set up as different partitions, but ultimately the choice of which region is written or read by the workload comes down to a simply key % num_partitions choice. While useful, it doesn't illustrate the power that could be shown by allowing the connection to determine which region to read/write from.