Nowadays Cloud environments are becoming a need for companies, research labs and also simple customers. Because of their several services, such a technology, has been succeeded to gather a high number of users around the world. One of the interesting services in Cloud computing is file storage, which consist on uploading our files to the Cloud's data centers and using them from anywhere at any time. Due to the highest number of users, an optimized management of the storage space around the data centers is required. Many studies have been focused on storage space optimization, researchers are trying to create some complex compression algorithms without considering their execution time and the user's behavior with a slow system. In this paper, we will try to save some storage space by classifying the customer files stored in the cloud's datacenter into three different clusters A, B and C, depending on the usage rate of each one. Cluster A represents the files which are the least frequently used. Cluster B represents the files which are the fair frequently used. And cluster C represents the files which are the high frequently used. Finally cluster A will be compressed by the algorithm which has the highest compression ratio and compression time results. And vice versa for cluster B. In the evaluation phase, our solution will be evaluated using a real dataset, and will be compared with some existing methods.