Some of the industrial processes with hard real-time constraints, such as those commonly found in the context of smart grids, require very reliable packet delivery and multicast. Solutions such as PRP [1] and iPRP [2] have been traditionally deployed over redundant (and often dedicated) networks. In this paper, we study how to construct node-redundant multicast trees that could be used in parallel over a shared network infrastructure. Our solution can be deployed over shared public networks, e.g., mobile carrier networks, hence it has an operating cost lower than traditional deployments of PRP and iPRP.We evaluate different algorithms proposed in the literature for providing node-redundant multicast-distribution-trees. We study specifically how to adapt these algorithms in an SDN network and compare them based on the number of forwarding rules that need to be installed on SDN switches, the number of hops between source-destination pairs given the installed forwarding rules, and the number of sources that can be placed in the network given the capacity constraints. In addition, we discuss the effects of topology changes (node failures, new source arrival and new destination arrival) on the activity of the SDN control plane.