Securing Cloud Data - Cloud Computing with e-Science Applications - page 66

3. Sentinel created: s Sentinels are created and appended to F ″ to yield F ′′′.
4. Permutation: The b ′ + s blocks of the file F ′′′ are permuted to yield F .
The prover (CSP) produces a concise proof that the archive retains and
reliably transmits the entire file or data object F. To ensure that the archive
has retained F, the verifier (client) V challenges the prover by specifying the
positions of a collection of sentinels in F and asking to return the asso-
ciated sentinel values. This phase includes Extract, Challenge, Respond,
and Verify functions. If the sentinels are returned correctly, then the file
has not been tampered with; if there are errors, then the error-correcting
code is used to retrieve the message. A drawback of this PoR scheme is the
preprocessing/encoding of F required prior to storage with the prover.
Shacham et al. [36] utilized two new economic and efficient homomorphic
authenticators. These authenticators are the primary encryption or hashing.
They also need larger storage requirements on the prover and provides proof
of security against impulsive adversaries.
Bowers et al. [7] introduced HAIL (high-availability and integrity layer),
a general conceptual framework for PoRs that is an improvement [20, 36].
It claims lower storage requirements and a higher level of security assurance
with minimal computational overhead and tolerates higher error rates than
scheme [20]. It is robust against an active, mobile adversary, that is, one that
may progressively corrupt the full set of servers. This work describes design
challenges encountered for practical implementation of PoR protocols. HAIL
is a distributed cryptanalytic system that allows a set of servers to prove
to a client that a stored file is intact and retrievable. Building blocks of the
HAIL system are the universal hash function, message authentication codes
(MACs), and integrity-protected error-correcting codes (IP-ECC). The advan-
tage of the HAIL adversary security model is that it ensures distributed file
system availability against a strong, mobile adversary.
The drawbacks of the PoR and PDP schemes are as follows:
• The effectiveness of these schemes rests primarily on the preprocess-
ing steps that the user conducts before outsourcing the data file. This
introduces significant computation and communication complexity.
• Most of these techniques do not support privacy preservation and
dynamic data operations.
• Most of these schemes focus on only static and archive data.
• None of these schemes considers batch auditing.
Public verifiability is needed in many cases when others should be able to
verify the data. A trusted TPA might have expertise and technical capabili-
ties that the clients do not have. Data audits by a trusted third party (TTP)
involve an independent authenticated entity to conduct a data audit.
Wang et al. [40] determined the difficulties and potential security issues
of direct extensions for fully dynamic data updates and then constructed