Hierarchical approach for green workload management in distributed data centers

The efficient management of geographically distributed data centers has become an important issue not only for big companies that own several sites, but also due to the emerging of inter-Cloud infrastructures that allow heterogeneous data centers to cooperate. These environments open unprecedented avenues for the support of a huge amount of workload, but they need the definition of novel algorithms and procedures for their management, where scalability is a priority. The complexity derives by the size of the system and by the need for accomplishing several and sometimes conflicting goals, among which: load balancing among multiple sites, prevention of risks, workload consolidation, and reduction of costs, consumed energy and carbon emissions. In this paper a hierarchical approach is presented, which preserves the autonomy of single data centers and at the same time allows for an integrated management of heterogeneous platforms. The framework is purposely generic but can be tailored to the specific requirements of single environments. Performances are analyzed for a specific Cloud infrastructure composed of four data centers.