Nowadays, storing and retrieving information over the Cloud is critical for the survival and growth of organizations and people. In this context, the possibility to store a huge amount of data and files on remote third-party Cloud storage providers is becoming an even more concrete practice. Unfortunately, there is not any guarantee regarding data availability and reliability of such providers. In such a context, if the Cloud wants to use the storage as a service of another Cloud, it has only to trust it. A possible solution consists of using a Multi-Cloud Storage (MCS) system. However, how organizations should compose their own MCS system on geo-distribution basis is not trivial at all. In this paper, we specifically discuss how to optimize the overall system in terms of data storage and retrieval by testing and validating a MCS system composed of three major Cloud Storage providers; Dropbox, Google Drive and Copy. Experiments have proved that the choice of the Cloud storage providers where to store files depends on the data transfer performance according to the file chunk size. In addition, we demonstrated that not always the provider that offers the best performance in upload also offers the best performance in download.

如今，通过云存储和检索信息对于组织和人员的生存与成长至关重要。在这种情况下，在远程第三方云存储提供商上存储大量数据和文件的可能性正在变得更加具体。不幸的是，不能保证此类提供商的数据可用性和可靠性。在这种情况下，如果云希望将存储用作另一云的服务，则它仅需信任它。可能的解决方案包括使用多云存储（MCS）系统。但是，组织如何在地理分布的基础上构建自己的MCS系统并不是一件容易的事。在本文中，我们将具体讨论如何通过测试和验证由三个主要的云存储提供商组成的MCS系统来优化整个系统的数据存储和检索。Dropbox，Google云端硬盘和副本。实验证明，云存储提供者在何处存储文件的选择取决于根据文件块大小的数据传输性能。此外，我们证明了并非总是在上传中提供最佳性能的提供者也会在下载中提供最佳性能。