To date the rapid growth of big data processing and its circulation among multiple organizations incur both promising prospects and security challenges for the corresponding technologies, such as data management, data analysis and so on. Efficient and secure data traceability is of critical importance for big data circulation, especially for cloud service applications which are not fully trusted and the risk of leakage of sensitive personal information. In this paper, we propose a framework for mutual traceability for data circulation and secure outsourced computation in data-centric cloud service. Our construction is built on top of searchable attribute-based proxy re-encryption. We enable both data owner and data user to trace their data circulation or perform privacy-preserving feedback. Specifically, the system enables data owners to efficiently distribute and trace his/her data to a specified group of cloud service providers who match a security/privacy policy and meanwhile, the data, maintaining its traceable property, can be further updated after being shared. The new mechanism is applicable to many real-world big data applications. Finally, our framework is proved chosen ciphertext secure in the random oracle model.