Modeling of resource monitoring in federated cloud using Colored Petri Net

In this paper, we have studied the problem of resource monitoring in federated cloud systems, which consist of multiple clouds. These systems are heterogeneous, large-scale, and distributed, and therefore, their management is very challenging. One important issue in these systems is to schedule the arrived tasks such that, the Service Level Agreement (SLA) requirements of the users are satisfied. For this purpose, the cloud systems should be aware of the resource status of each other. Each cloud should update its information about the workloads and resources of other clouds, to be capable of performing task scheduling in an efficient manner. To tackle this problem, we have proposed a low-overhead mechanism for communication among the clouds. This approach exploits both of pull and push strategies, which yields in providing up to date information with low overhead. We have also modeled the proposed algorithm by Colored Petri Net. The model is implemented and analyzed using CPN-tools.