FE4MQTT - Using Functional Encryption to Improve the Privacy in Publish-Subscribe Communication Schemes

Publish-subscribe communication schemes, such as the lightweight Message Queue Telemetry Transport (MQTT) standard, where a central broker forwards messages between clients, may have certain benefits compared to direct communication channels. However, a fundamental problem in such schemes is, that the broker can observe and analyze the message exchange between two or more clients. This way, he might get privacy related insights about a clients’ behavior. In contrast to the messages’ payload, classical encryption can not be used to protect a message’s topic, as the broker must be able to interpret it. In this paper, we propose the use of Functional Encryption (FE), an advanced public-key encryption scheme, to overcome this problem. With FE, the broker can use a ciphertext as input to execute a function. In our experimental setup, a MQTT broker will not able to learn anything about a message type (i.e. topic), but is still able to forward the messages to the correct subscribers. Furthermore, the approach allows using a randomized message topics during publication for a single message type. The proposed concept will be evaluated with regard to the computational overhead and scalability for the use in an IoT environment.