|
Proposals | Key considerations |
Reliability and quality | Reputation and trust | Tolerance and security | Consensus |
|
He et al. [5] | Similarity criteria are addressed via a smart contract algorithm | No reputation or trust criteria are defined | Tolerance and security topics are not addressed | No specific consensus mechanism is referenced |
Gong and Lee [6] | Abstract data verification mechanism | Abstract evaluation of contributors credibility | Presents a mechanism to prevent sybil attacks from malicious contributors | Abstract miner-based consensus mechanism |
Mendez Mena and Yang [7] | No quality criteria are defined | Trust is established only via utilising a permissioned block chain environment | Tolerance and security topics are not addressed | Proof of authority |
Cha et al. [8] | CTI data are subject to verification via a collaborative architecture but is not clear under which criteria | No reference to reputation criteria or supported mechanisms | No reference to tolerance against malicious activities | No reference to consensus algorithms |
Meier et al. [11] | The quality evaluation is based on correlation and contribution graphs with no detailed quality criteria | No reputation or trust metrics are addressed, since this proposal does not require a ground truth | Robust against a small percentage of dishonest contributors but susceptible to malicious attempts of a larger percentage | Not applicable |
Riesco et al. [9] | Quality criteria of identity, authority, motive, access, timeliness, and consistency are defined | Trust is presented as an overall benefit using a blockchain-based network, but there is no reference of quantification of reputation and/or trust between peers; a Cobb–Douglas utility function is presented combining trust and quality | Inherited by the block chain technology with no specific reference | No reference to consensus algorithms |
Wu et al. [12] | Quality criteria of completeness, freshness and relevance are applied but the assessment is performed under a vague methodology | The reputation assessment is using EigenTrust algorithm [14] to calculate a reputation score based on peer transaction histories and produce global trust values for all participants | No reference to tolerance against malicious activities | It is inferred that the consensus algorithm of proof of elapsed time (PoET) is proposed |
Proposed solution | The model depends on literature-referenced quality criteria evaluation with explicit methodology, metrics and indicators | Trust is created via specific mechanisms and processes as well as the reputation is derived from the data stored on the ledger as a historical immutable reference | The theoretical simulation showed that there is tolerance against malicious validators even if the ratio of legitimate vs malicious validators is 1 : 50 | The model proposes a new consensus algorithm, namely, proof-of-quality which is based on a voting procedure among the best performant validators |
|