Fig 1.
DIA-MTTP high-level architecture.
Table 1.
Notations used in the design of DIA-MTTP.
Table 2.
Math equations for tag generation.
Fig 2.
Distributed M2T data structure.
Fig 3.
DIA-MTTP: Functional blocks, components and entities.
Fig 4.
The D3U architecture.
Fig 5.
The D3U protocol suite.
Fig 6.
D3U protocol: Sub-protocols and their messages.
Fig 7.
The LoA1DV architecture.
Table 3.
Math equations for public verification.
Fig 8.
The LoA1DV protocol suite.
Fig 9.
LoA1DV protocol: Sub-protocols and their messages.
Fig 10.
The LoA2DV architecture.
Table 4.
Math equations for private verification.
Fig 11.
The LoA2DV protocol suite.
Fig 12.
LoA2DV protocol: Sub-protocols and their messages.
Fig 13.
The DU architecture.
Fig 14.
The DU protocol suite.
Fig 15.
DU protocol: Sub-protocols and their messages.
Table 5.
Cryptographic operations and their computational time (in seconds).
Fig 16.
Data corruption probability vs the number of requested blocks under different detection probabilities.
Table 6.
Number of encrypted data blocks and their associated tags: With/without the data deduplication approach.
Fig 17.
Computational cost of the user in the D3U vs the number of data blocks.
Fig 18.
Computational cost of the user in D3U vs the number of redundant data blocks (K = 1000, using Exp2).
Fig 19.
Computational cost of the leader provider in the D3U vs the total number of data blocks in the file.
Fig 20.
Computational cost of the leader provider in D3U: With/without the data deduplication approach.
Fig 21.
Computational cost of the providers (Leader and non leaders) in D3U: With/without the hierarchical approach.
(K = 1000 data blocks).
Fig 22.
Communication cost of the user in D3U: With/without data deduplication.
Fig 23.
Communication cost of the user in D3U: Hierarchical approach vs non-hierarchical approach.
(K = 1000).
Fig 24.
Communication costs for the providers vs the number of data blocks: With/without a hierarchical approach.
(n = 6 PCSes, * is Non-Hierarchical approach).
Fig 25.
Communication cost incurred by the leader provider regarding the number of data blocks and PCSes.
Fig 26.
Computational cost of PCS in LoA1DV: With/without nonces.
Fig 27.
Communication cost of the L-TPA in LoA1DV vs the TPAs and the number of data blocks.
(|En_DB| = 0.025 KB and |DBTag| = 0.032 KB).
Fig 28.
Communication cost of the L-TPA in LoA1DV against the number of data blocks.
(* is Key-based approach, n = 2).
Fig 29.
Computational costs for the leader provider and the non leader provider in LoA1DV and LoA2DV vs the number of data blocks.
(n = 20).
Fig 30.
Communication cost for the user in LoA1DV and LoA2DV.
(* the NonKey-based approach, and C = 20 data blocks).
Fig 31.
Communication cost for the leader provider and the non leader provider in LoA1DV and LoA2DV.
(C = 100, n = 2, |m| = |G1| = 0.032 KB, considering only the proofs cost).
Fig 32.
Communication cost of the non leader provider in LoA1DV and LoA2DV vs the data block number.
Table 7.
Computational cost of tag generation against different sizes of data file and data block (in seconds).
Fig 33.
Communication costs of the L-TPA and the non leader TPA in LoA1DV and LoA2DV.
(n = 10, |m| = |G1| = 256 bits, |p| = 200 bits, using Key-based approach).
Fig 34.
Computational cost of the user in DU vs the number of data blocks.
Fig 35.
Storage cost at the TPAs: With/without the collaborative verification approach.
(10 TPAs, |n2| = 0.256 KB).
Fig 36.
Storage cost for the DIA-MTTP entities with a different number of data blocks.
(10 TPAs, |m| = |G1| = 0.032 KB, |n2| = 0.256 KB, * without the collaborative verification approach).
Fig 37.
Storage cost for each entity in the DIA-MTTP: With/without data deduplication.
(The redundancy data rate 20%).
Table 8.
Comparing the DIA-MTTP with existing DIAs against the functional, security and reliability requirements in requirement specification section.
Table 9.
Comparing the DIA-MTTP with existing works against the efficiency requirements in section.