Combining models of capacity supply to handle volatile demand: The economic impact of surplus capacity in cloud service environments

In the paper at hand we analyze the capacity planning problem of a service vendor providing a business process characterized by volatile demand to his customers. Thereby, we consider the situation that the service vendor executes certain activities by himself whereas specific parts of the business process are outsourced to external providers. For the outsourced parts, the vendor can choose between different models of capacity supply (MCS) that are offered by external providers differentiating with respect to elasticity of provided capacity and the underlying pricing model. Thereby, in addition to the two “traditional” MCS dedicated capacity and elastic capacity, recent developments in information technology enable the on-demand use of surplus capacity from an external providers' market. Since an integrated analysis of these three MCS is still missing in common literature, we develop an optimization model allowing for the simultaneous consideration of the three different MCS within an integrated queuing system. By analyzing the optimization model with help of a discrete event simulation, we study the question of how these different MCS may be combined to minimize the total operating costs of the service vendor considering volatile demand. The simulation results show that combining different MCS tends to be favorable in contrast to the stand-alone usage of a certain MCS. In particular, combining the additional option of using surplus capacity with “traditional” MCS promises cost advantages. Our optimization model therewith provides first insights in the potential economic benefits of IT-enabled MCS.