Demand heterogeneity in IT infrastructure services: Modeling and evaluation of a dynamic approach to defining service levels

A key feature of service-oriented models of information technology is the promise of prespecified quality levels enforceable via service level agreements (SLAs). This poses difficult management problems when considerable variability exists in user preferences and service demand within any organization. Because variance in expectations impact service levels, effective pricing and resource allocation mechanisms are needed to deliver services at the promised quality level. In this paper, we propose a mechanism for SLA formulation that is responsive to demand fluctuations and user preference variance, with the objective of maximizing organizational welfare of the participants. This formulation features a dynamic priority based price-penalty scheme targeted to individual users. An analytical model is presented and evaluated for effectiveness of a proposed dynamic priority-based pricing scheme vis-a-vis a baseline fixed-price single-quality level SLA. Simulations using data from an existing SLA is used to provide evidence that the proposed dynamic pricing scheme is likely to be more effective than a fixed-price approach from a system welfare perspective.