SLA-Based Trust Model to Enhance Availability in Private Cloud

Abstract

To solve the problem of trust in cloud, various cloud computing trust models has been proposed in the literature. Among those Service Level Agreement (SLA) based trust model is one.  Lack of reliability and availability have always been a major concerns in cloud computing. Research reports express that about $285 million have been lost yearly due to cloud service failures and offering availability of about 99.91%. Replication has been widely used as a means of increasing the data availability of large-scale cloud storage systems where failures are normal. However, replica placement in cloud heterogeneous storage system becomes more complicated when compared to a small-scale system, where each data node may have different capability and can only admit a restricted number of requests. Hence, where to keep replicas in the Data Center to satisfy availability requirement is a matter of concern.  Aiming at the problem in this paper a trust model based on Service Level Agreement (SLA) is proposed.  In this model, when the cloud node service provider (Data Center Node) registers with the service center—a trusted third-party platform, it first submits its own strength evaluation report, stating its strengths. The service center based on report evaluates the cloud node service provider’s system trust. Secondly, the system trust is combined with direct trust and indirect trust to calculate comprehensive trust of the cloud node service provider. Finally, the user negotiates the SLA with the cloud node service provider based on the service and comprehensive trust to determine the node for replication. The experimental results show that the proposed trust model based on SLA improves the success rate of node selection for better availability.