Much research has focused during the last years on the security and privacy concerns of public cloud storages. Cryptographic primitives are commonly used to ensure user data confidentiality, authenticity and integrity. Confidentiality has been addressed by the use of symmetric-key encryption algorithms, while integrity and authenticity have been achieved by using message authentication codes, secure hashes or digital signatures. The choice of a specific configuration for securing an untrusted cloud storage highly depends on the expected security level, the size and type of data to store and the access pattern to these data. In this work, we are interested in overcoming the lack of comprehensive comparison of the costs and effectiveness of cryptographic primitives for securing public cloud storage, and ease an informed choice between them based on target usage conditions. We describe the results of an independent experimental study of six cryptographic schemes, representative of the principal design alternatives. Our practical experience report reveals that the best scheme for a given situation, such as a write-heavy workload of mostly small files, is not necessarily the most appropriate for a different situation such as a read-only workload of large files. We identify the scheme characteristics that are correlated with these differences and discuss the pros and cons of each design. Our experimental framework and results are available in the open for use by the community.