Currently a new worker with a new swarm instance is spun up for each request. This is like running multiple clients to execute requests in parallel. We should be running requests in parallel from a single client. This means a single background thread should spin up a child worker for each transfer.
Currently a new worker with a new swarm instance is spun up for each request. This is like running multiple clients to execute requests in parallel. We should be running requests in parallel from a single client. This means a single background thread should spin up a child worker for each transfer.