Skip to main content

Hadoop Characterization

Characterization of hadoop networking traffic towards programming of the network resource allocations.

Network characterization of Hadoop

We have instantiated a Hadoop workload on GENI to analyze the networking requirements of a typical Hadoop job.

Our purpose is to investigate networking requirements and then come up with an optimized networking setup for a given Hadoop job. This was demonstrated during the GEC19 demo session.


OpenFlow flow rules were identified to characterize the network usage and behavior of Hadoop.

Network characterization of Hadoop

Our characterization work continued by identification of all protocols utilized by Hadoop during GEC20 demo.