ESCM: A Toolkit for Efficient and Secure Outsourced Computation with Multiple Keys

With the widespread availability of cloud computing and big data, homomorphic encryption has become a crucial technique employed to protect the data privacy in outsourced computation. However, the existing outsourced computation toolkits based on homomorphic encryption either require all participants to share the same key or result in enormous computational and communication overheads. In order to address these shortcomings, we put forward a toolkit for efficient and secure outsourced computation in multiple key scenarios (ESCM). ESCM can permit the servers to process the most frequently used arithmetic operations such as multiplication, division, sorting and so on across different encrypted domains. Moreover, to tackle the security concerns that may arise from collusion among some servers, as well as the problem of service disruption due to server outages, we propose the distributed two trapdoor cryptosystem with threshold decryption, the core cryptographic primitive, which is able to support (k, n) threshold decryption. Theoretical analysis validates the security of the proposed ESCM and compares the computation and communication complexity with existing most advanced solutions. Finally, simulation experiments illustrate the practicality and efficiency of ESCM.