A Coloured Petri Net Model for Hybrid NFV-MEC Systems Supporting URLLC

Integrating Multi-Access Edge Computing (MEC) and Network Function Virtualization (NFV) favors URLLC, but introduces challenges in resource allocation due to the reduced dimension of NFV-MEC nodes, the overhead caused by virtualization, the use of heterogeneous virtualization technologies, and potential failures during service processing. These factors require careful consideration in resource dimensioning/allocation to ensure that URLLC requirements are met without incurring excessive power consumption due to overprovisioning. This paper proposes a Coloured Petri Net (CPN)-based model for dynamic resource allocation in NFV-MEC systems supporting URLLC services. The proposed model represents a hybrid NFV-MEC system, accounting for virtualization overhead, failure costs, service buffering, and dynamic resource scaling based on demand. It also accommodates system features with diverse behaviors and provides analysis of the system's availability, latency, and power consumption. Results showed that for high arrival rates, improving the container service rate has a greater impact than adding VMs. The benefits of a higher container setup rate on response time become noticeable when VM capacity is exceeded.