LPN-based authentication protocols can be very efficient. This makes them a very good choice for certain applications, for example in the RFID setting. A major concern in this kind of setting is information leakage due to side-channel attacks. In practice, an adversary could be able to extract information about secret values from the computation performed in the protocol.

In this talk, I will present ongoing work about a framework to make existing LPN-based protocols robust against information leakage with a bounded computational cost. The framework makes use of a technique known as "masking." Our instantiation of the inner-product masking guarantees security in the presence of leakage in a reasonable and widely accepted security model.