From: TMaR: a two-stage MapReduce scheduler for heterogeneous environments
 | Workload type | Input data size small, medium, large | # of Map Tasks |
---|---|---|---|
Scenario 1 | Job is CPU-intensive and the produced intermediate data is high in the range of [30, 300] GB | <1 GB, 3 GB, 5 GB, 10 GB> | <8, 24, 40, 80> |
Scenario 2 | Job is CPU-intensive and the produced intermediate data is low in the range of [10, 30] GB | <1 GB, 3 GB, 5 GB, 10 GB> | <8, 24, 40, 80> |
Scenario 3 | Job is IO-intensive and the produced intermediate data is high in the range of [30, 300] GB | <1 GB, 3 GB, 5 GB, 10 GB> | <8, 24, 40, 80> |
Scenario 4 | Job is IO-intensive and the produced intermediate data is low in the range of [10, 30] GB | <1 GB, 3 GB, 5 GB, 10 GB> | <8, 24, 40, 80> |