There is unused code for using parallelCollectionScan in the CollectionCloner for initial sync. This was possible using the AsyncResultsMerger, which supports sending getMore commands for multiple cursors and processing the results asynchronously. However, for the Faster Initial Sync project, we are rewriting the CollectionCloner to use DBClientConnection. In order to support multiple cursors with DBClientConnection, we would need to create multiple data cloning threads. This is not worthwhile, since we are not currently using this functionality. WiredTiger does not support parallelCollectionScan and has no plans to do so.