Increasing the efficiency of workload management in data
centers is essential to achieve several business and technical
goals, such as reduction of costs, energy consumption and
carbon emissions. Many solutions are available for workload
management in a single data center, but there is still much
space for the development of frameworks that are able to
manage the workload in a distributed scenario, with multiple
sites and data centers. In this paper, we present the main
benefits of hierarchical solutions, in which the problem is decomposed
into two layers, a lower layer that focuses on the
workload management within each single data center, and an
upper layer that orchestrates the management of the workload
on a multi-site environment. We focus on the main advantages
of hierarchical approaches, i.e., autonomous management,
scalability and modularity, and illustrate how and
to which extent these advantages can be exploited in some
emerging business scenarios, i.e., geographical data centers,
software data centers and Hybrid Cloud.
