Differentially Private Billing with Rebates

  • George Danezis ,
  • Markulf Kohlweiss ,
  • Alfredo Rial

MSR-TR-2011-10 |

A number of established and novel business models are based on fine grained billing, including pay-per-view, mobile messaging, voice calls, pay-as-you-drive insurance, smart metering for utility provision, private computing clouds and hosted services. These models apply fine-grained tariffs dependent on time-of-use or place of-use to readings to compute a bill.

We extend previously proposed billing protocols to strength\-en their privacy in two key ways. First, we study the monetary amount a customer should add to their bill in order to provably hide their activities, within the differential privacy framework. Second, we propose a cryptographic protocol for oblivious billing that ensures any additional expenditure, aimed at protecting privacy, can be tracked and reclaimed in the future, thus minimising its cost. Our proposals can be used together or separately and are backed by provable guarantees of security.