A protocol for Mixed-Criticality management in Switched Ethernet networks

In real-time industrial networks, providing timing guarantees for applications of different criticalities often results in building separate physical infrastructures for each type of network at the price of cost, weight and energy consumption. Mixed-Criticality (MC) is a solution first introduced in embedded systems to execute applications of different criticality on the same platform. In order to apply MC scheduling to off-the-shelves Switched Ethernet networks, the key issue is to manage the criticality change information at the network level. The objective of this work is to propose a criticality change protocol for MC applications communicating over Switched Ethernet networks. The protocol relies on a global clock synchronization, as provided by the IEEE-1588v2 protocol, and a real-time multicast based upon it, to preserve the consistency of the criticality level information stored in all the Ethernet switches. We characterize the worst-case delay of a criticality change in the network. Simulation results show how the criticality change delay evolves as a function of the network load.