We model a queueing resource management in mutual mode. When analyzing the performance of the service system, we often use the queueing theory to get the various performance indexes. As the number of resources in the service system with interactive mode increases due to users' sharing behavior, it becomes very difficult to get a steady state of the system. We investigate how to evaluate the performance of such a service system in which the amount of resources changes over time on accounts of users' sharing behavior. In the process of solving, an algorithm is designed to evaluate the convergence of performance indicators. The algorithm effectiveness is verified by the simulation test of the evaluation system and the simulation frequency range of convergence of service performance indicators. We compare the queueing process in the mutual mode to the non-mutual mode and then obtain a steady state for the mutual mode. The unique phenomenon in interactive mode, which is a performance inflection point, is found which has a guiding significance for efficient production.