| Abstrakt: |
Cloud computing is arising as a popular computing paradigm, as it is good in offering its users on-demand scalable resource-based services over the internet. In peak hours, a single cloud is not efficient in serving an application; therefore, the collaborative cloud model has been introduced. Collaborative cloud computing (CCC) uses the globally-scattered distributed cloud resources of the diverse organizations collectively in a cooperative manner to provide the required service to the user. The allocation, as well as the management of the resources, is being a challenging task in the CCC, due to the heterogeneity of the resources. On the other hand, the assurance of the Quality of Service (QoS) and reliability of these resources is challenging. Further, it would be efficient if the resources are provided based on the system behavior. In this research work, a novel trust computing model is developed, which predicts both the QoS and Trust via analyzing the system behavior. The proposed model encloses three major phases: Trust-QoS behavior estimation, resource matching and resource allocation. Initially, the QoS and Trust behavior of the system is estimated via a Neural Network (NN) model. Subsequently, the resource allocation is performed using the parallel resource matching framework, which is based on the concept of Map-Reduce. More particularly, the precious resource allocation is achieved by an optimization logic called Improved Grey Wolf Optimizer (I-GWO). Here, the improvement of GWO emphasis the consideration of both the best and worst fitness. [ABSTRACT FROM AUTHOR] |
