Is your feature request related to a problem? Please describe.

Currently, the S3 storage option for the PMGD based query handler is very slow, due to the fact that each query is forced to instantiate an entirely new connection object.

Describe the solution you'd like

The S3 connections can be natively multithreaded with no additional overhead. We should move to a "shared connection" model so that the server process can repeatedly re-use the same connection that is instantiated at server start.