Improving QoS assessment over multiple attributes with asynchronous data

摘要

When multiple providers offer functionally identical services in a SOC environment, it is only natural that consumers should ask how their qualities would compare. While various methods have been proposed to help assess the quality of a service (QoS) using monitored quality data, they do not handle multiple QoS attributes adequately, especially when the qualities of these attributes are monitored asynchronously. In this paper, we propose a method that takes both accuracy and confidence into account when assessing QoS over multiple attributes, and employs a &NN based technique to deal with asynchronous data. Our experiments show that the new method can give more effective QoS assessment over multiple attributes than the existing methods do.