On the Resource Consumption of Secure Data Sharing

Abstract

Sharing data within a closed user group is the basis for lots of applications today in both personal and professional life. The risk that the storage provider or an attacker illegitimately inspects or tampers with the shared data can be mitigated by employing client-side cryptography. In this work, we estimate the resource consumption that secure data sharing based on client-side cryptography requires in terms of computation time and network traffic volume. We therefore go beyond asymptotical analyses and state the absolute resource consumption for different secure data sharing protocols and client devices under realistic sharing models. These sharing models are extracted from traces of real-world collaboration platforms running in production. To the best of our knowledge, this is the first time that such a characterization of sharing models is published. Furthermore, we provide a clearly defined resource consumption estimation model. Our results show that for users who just up-and download data, the execution time of required cryptographic operations is typically up to a few seconds. The results also show that group owners have to deal with significantly higher computation times and network traffic when a user is removed from a sharing group with a few hundred or more members, given that basic secure data sharing protocols are used as they are in place today. A further finding is that computation time and network traffic volume are considerably lowered by extending the secure data sharing protocol with group key management approaches, at the price of slightly raised computation times for smaller groups.