Secure Cloud-based Audio Storage and Processing
A client/customer who is constrained in storage wants to securely outsource storage of confidential audio secret to CDC. An important aspect of our outsourcing model for storage is that the CDC does not learn any information about the confidential audio data. We apply (K,N) SSS to encrypt the audio secret into N shares that can be distributed among N CDCs (non-colluding) such that at least K out of N number of shares can be retrieved by an authorized user to reconstruct the secret; individual shares are of no use on their own. Our proposed scheme solves the limitations of previous Audio Secret Sharing techniques which:
(i) does not extend (K,N) threshold
(ii) security of the scheme is not proven
(iii) suffers from the limitations of using Human Auditory System (HAS) to decrypt the secret.
Below figure illustrates our proposed method for secure audio storage.
Secure Cloud-based Audio Storage
Below is the audio secret, it's shares and reconstructed audio secret of a sample audio file from our dataset after applying our proposed method. Listening to the files reveals that (i) the share is noise and does reveal information about the audio secret perceptually and (ii) the reconstructed audio secret sounds identical to the original secret. Security and performance analysis suggest that: (i) our scheme is information theoretically secure meaning that an adversary with unlimited computing power cannot obtain any information about the secret, (ii) the transmission overhead of a share from client to CDC is not high and (iii) the audio secret is reconstructed with minimal losses as a result of round-off errors.
