.png){kind=icon}
.png){kind=icon}
Return
Reputation based novel trust management framework with enhanced availability for cloud
.png){kind=icon}
PRE
AI
DOI：10.1016/j.jpdc.2023.03.010.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
To select a cloud service provider (CSP) out of immense number of relevant CSPs available, cloud customer (CC) trusts on the reviews provided by the CCs who have used the services provided by CSPs. However, relying on these feedbacks may be hazardous if the review(s) have been deliberately posted with wicked motive. To mitigate the impact of malicious feedbacks, a framework, namely highly available credible cloud trust management framework (HA2CTMF) has been proposed with the intent of revamping the reputation and availability parameters of QoS by evaluating the trust of cloud service providers (CSPs) based on the credibility of feedbacks furnished by CC. HA2CTMF detects collusion and Sybil attacks on the labeled CloudArmor dataset and increases availability of the trust service. Collusion attack has been detected by comparing review count; service count; & mean of reviews with their respective threshold values. Further, Sybil attack has been uncovered with the help of these aspects: deviation from mean of reviews; pattern detection; & comparison of polarity value from both lower & upper threshold values. Additionally, enhanced availability has been ensured by working on disaster management and diminishing repercussions of single point of failure. The results have been accessed on the following parameters: precision; recall; f1-score & availability percentage on the dataset comprising 8774 data points. This framework has exhibited precision of 85.69%, recall of 89.45%, f1-score of 87.53% and availability of 96.71% compared to CloudArmor framework values on the same parameters as: 44.47%; 62.21%; 51.85% & 70.43% respectively. As per the results attained, we conclude that proposed HA2CTMF has demonstrated better outcome.& COPY; 2023 Elsevier Inc. All rights reserved.
Keywords:
Reputation management
Trust management model
QoS performance
Fake feedback detection
Cloud customer reviews
Journal
IF:
4
Papers:
3.8K
Citations:
4.8K