Sustainable dynamic application hosting across geographically distributed data centers

Geographical appLication/WorkLoad Placement (GLP) shifts workload toward data centers that offer lower electricity price or green energy at a given time. Despite being advantageous, it is challenging mostly because of (i) applications' migration overhead, (ii) prediction errors of workload and renewable over a large prediction window, (iii) the non energy proportionality of servers, and (iv) the physical limitations of batteries. This chapter accounts for general problem statement of GLP through developing a cost optimization problem and discusses the challenges and existing solutions. Further, we perform a simulation study to show the effectiveness of GLP towards sustainability of data centers.