From: A comparison on scalability for batch big data processing on Apache Spark and Apache Flink
Dataset
Spark MLlib
Flink
Difference
ECBDL14-10
42
111
69
ECBDL14-30
61
196
135
ECBDL14-50
103
302
199
ECBDL14-75
123
456
333
ECBDL14-100
174
783
609