parallelising query execution

[rcelebi]

If the federated query retrieves large intermediate results due to request split, it can take a long time. To mitigate this, you can download the whole intermediate results once and then join them