Stadam: A new SLA trust model based on anomaly detection and multi-cloud
Yu Jin, Shaoying Min · 2016
Cloud computing is a new commercial paradigm. It brings opportunities as well as challenges for cloud users, especially for small and medium size IT business. Trust is one of the important issues. In order to receive the high quality services, cloud users must build trust in the cloud providers. Trust model is one of the effective solutions. However, the existed trust models have defects: ©most of the models cannot evaluate the trust in the cloud provider when service is running; ©Although very few of models can do this, they have a common disadvantage: service is provided by single cloud. In other words, it is incapable of seamless shifting from the running service against Service level Agreement (SLA for short) to the other service. So there is a short time when the service is unavailable. Point to this, Stadam is proposed in this paper which is a new SLA trust model based on the anomaly detection and Multi-cloud. In Stadam, cloud consumer synchronously uses multiple services that are real-time monitored if SLA is fulfilled by using anomaly detection algorithm. Then the SLA trust is updated. According to the current trust the number of requests that are sent to service is dynamically changed. So Stadam not only detect abnormal ongoing service, but also seamlessly switch between multiple services.