Anonymous Stochastic Routing

We propose and analyze a recipient-anonymous stochastic routing model to study fundamental trade-off between anonymity delay. An agent wants quickly reach goal vertex in network through sequence of actions, whereas an overseeing adversary observes the agent’s entire trajectory tries identify among those vertices traversed. are interested understanding probability that can correctly (anonymity) as function time it takes (delay). A key feature our is presence intrinsic uncertainty environment, so each intended steps subject random perturbation thus may not materialize planned. Using large-network asymptotics, main results provide near-optimal characterization anonymity–delay under number topologies. Our technical contributions centered on new class “noise-harnessing” strategies adaptively combine from environment with additional artificial randomization achieve provably efficient obfuscation.