The dynamic nature of power systems combined with the need for low-latency and loss-tolerant communications, presents significant challenges to maintaining system reliability and resiliency. This paper proposes a novel integration of Finite Control Set Model-based Predictive Control with an extended prediction horizon and Software Defined Networked to address the resiliency problem and voltage/frequency deviations associated with traditional hierarchical microgrid. The communication framework integrates Software Defined Networked as a set of microservices distributed across local controllers and improved system reliability under communication constraints. The secondary control considers the variability of communication latency and packet loss to adjust the shared reference based on the spatial and temporal correlation. The microgrid is subjected to four test scenarios to analyze the impact of communications on distributed generation, plug-and- play capacity and load variations. The proposed control framework significantly improves system performance, achieving a 0.2-0.3 s recovery time, 0.05 s communication latency, and maintaining stability with up to 60% packet loss. Compared to hierarchical methods, it reduces recovery time by up to 90%, frequency deviation by up to 80%, and enhances power sharing and coordination between distributed generators. This method addresses the problem of low dynamic response of control strategies during disturbances, allowing the implementation of new, reliable and resilient hierarchical microgrids.