| The development of big data,cloud computing and artificial intelligence increases the popularity of data sharing and its applications.Meanwhile,new challenges also emerge due to the problems of data abuse and privacy leakage in data sharing platforms today.The main goal of our research is to provide two solutions for limited privacy protection in a data sharing platform called Data Mesh using the trusted execute environment(TEE)and secure multi-party computing(SMPC)respectively.The first solution based on the trusted execute environment can increase the privacy protection ability of the model interpreter of Data Mesh.And the use of intel remote attestation brings the immutability of data sharing models.The other solution under the secure multi-party computing protocol makes it possible for multi parties to share data without retrieving the source data from others.We experimented the effectiveness of these two solutions and compared their pros and cons.The solution built on TEE achieved better performance,while the other one using SMPC software still worked well in machines with basic settings.The future research can focus on combining these two solutions to create a data-sharing as a service platform with strong privacy protection ability and personal customization. |
