feature/#19 added perf benchmark for creating pipeline, task and dn entities

[toan-quach]

19

@jrobinAV

Some of the things we can improve further from this draft:

• the perf benchmark for entities (scenario, pipeline, task, datanode) creation can use a general abstract class similar to the class DataPerfBenchmark for DataNode performances
• reset config, scheduler, etc. after each run
• the config for using either SQL or FS repo can be done in _generate_configs