Techniques for data routing and management using risk classification and data sampling

A computer-implemented method, comprising:

• generating a traffic sample from data transiting from a source to a destination on a network;
• generating a replicated traffic sample from the traffic sample, the size of the traffic sample adjusted based at least in part on a risk profile associated with the data;
• processing the replicated traffic sample by at least comparing an expected behavior of the replicated traffic sample with an observed behavior of the replicated traffic sample and based at least in part on a characteristic of the destination;
• and
• initiating a mitigation measure for the data based at least in part on the observed behavior differing from the expected behavior.

A system comprising:

• at least one physical computing device that implements one or more services, wherein the one or more services:
• generate a replicated traffic sample from a traffic sample;
• generate a plurality of risk level components based at least in part on the replicated traffic sample;
• compare an expected behavior of the replicated traffic sample to an observed behavior of the replicated traffic sample;
• and
• initiate a mitigation measure for data, associated with the replicated traffic sample based at least in part on the observed behavior differing from the expected behavior.