Implementation and evaluation of publicly verifiable proofs of data replication and retrievability for cloud storage

Abstract

Cloud storage is a mature and widely used cloud technology. Behind this, it involves the proofs of the retrievability of files stored on the server, which allows the client to remotely check whether its files are correctly stored on the cloud server. In 2020, Gritti proposed the first publicly verifiable Proofs of Retrievability and Reliability (P-PORR). P-PORR combined Proofs of Retrievability (POR) with Verifiable Delay Functions (VDF) providing a fast verification and allowing anyone to verify the server’s behavior, not just the client. We built a realistic cloud test environment, implemented and tested P- PORR in this environment. This paper describes the implementation process, analyzes the performance of P-PORR from multiple perspectives, such as the total time spent before files are uploaded to the server, verification time and financial considerations. We stand in the perspective of the client and server to discuss the effect of the results. The results show that P-PORR has a similar performance to another PORR, with private verification, called Mirror. It even outperforms Mirror in processing small files. Therefore, P-PORR is suitable for cloud storage services. In the long run, this paper provides a practical application instance of VDF and a performance comparison benchmark for later PORR research.