Service Level Agreements (SLA) for cloud services entail multiple, interrelated variables such as price, penalty, contract length, and service availability (uptime) guarantee, representing key trade-offs faced by the service provider. In this study, we address two significant components of the SLA-based contract design problem: (1) the availability-aware service provisioning problem to minimize the expected total cost and (2) the pricing problem for a given penalty level. We present a binary-search algorithm to derive the optimal number of backup resources to manage availability. We also derive closed form expressions for the minimal price to charge for the service. In addition, we design a periodic intervention strategy that allows the provider to adjust the allocation of backup resources such that the expected total cost is minimized. We also present some preliminary computational results exploring the sensitivity of our solutions (e.g. service provisioning and pricing) to certain parameters in our contract model.
Session 3B: Cloud Computing Paper 2