Pricing Strategies of Public Cloud Computing Instances: A Game-Theoretic Model

Public cloud capacity providers offer various pricing models namely, on-demand capacity instances, reserved capacity instances, and a hybrid of these two pricing models. The buyer firms seeking cloud solutions not only source their requirements from the public cloud providers but also invest in some private in-house information technology (IT) infrastructure. In this paper, we find conditions for implementing each of the above pricing models by a public cloud provider given that the buyer firm will also invest in a private cloud. We develop a game-theoretic model to determine the cloud provider's pricing schemes and the conditions to offer different pricing models. We discuss various insights and important business implications for public cloud providers. For example, we find that if the available capacity at the public cloud provider's end is high, the prices in both the reserved capacity instances model and the on-demand instances model are low. Interestingly, and somewhat counter-intuitively, in the hybrid pricing model, we find that if the available capacity is high, the unit subscription price is low but the unit on-demand price is high. We also observe that when the cloud provider's capacity is low, it is optimal for the provider to implement a pure on-demand pricing model. Further, in the intermediate range of capacity, the public cloud provider should implement a hybrid pricing model. However, at a high range of capacity, the public cloud provider should prefer a pure reserved capacity instances model. In addition, we provide other interesting and useful insights for the public cloud capacity providers that may help them in formulating their pricing strategy.