Collaborative filtering (CF) systems are being widely used in E-commerce applications to provide recommenda- tions to users regarding products that might be of interest to them. The prediction accuracy of these systems is depen- dent on the size and accuracy of the data provided by users. However, the lack of sufficient guidelines governing the use and distribution of user data raises concerns over individ- ual privacy. Users often provide the minimal information that is required for accessing these E-commerce services. In this paper, we propose a framework for obfuscating sen- sitive information in such a way that it protects individual privacy and also preserves the information content required for collaborative filtering. An experimental evaluation of the performance of different CF systems on the obfuscated data proves that the proposed technique for privacy preser- vation does not impact the accuracy of the predictions. The proposed framework also makes it possible for mul- tiple E-commerce sites to share data in a privacy preserving manner. Problems such as the cold-start scenario faced by new E-commerce vendors, and biased results due to insuf- ficient users, are resolved by using a shared CF server. We describe a centralized CF server model in which a central- ized CF server makes recommendations by consolidating the information received from multiple sources.