Cloud computing often uses a multi-tenant architecture where tenants share application and system software. Request prioritization presents a challenge in this architecture. Tenant may have individual (local) prioritization requirements, and these requirements can be different for different tenants. The shared application must use a global priority scheme for requests from all the tenants. This paper proposes an effective model to prioritize service requests from multiple tenants while preserving local priorities from individual tenant requests. This paper proposes the Crystalline Mapping (CM) algorithm which maps local priorities from individual tenants to global priorities. The algorithm also maximizes revenues within the local-to-global priority mapping constraints. The performance of the model is evaluated by the total business value, the QoS measurement, and a fairness measure. The model has been evaluated using simulation based on real-world data, and the results are consistent with the mathematical model.