Mission-critical applications that use computer networks require guaranteed communication with bounds on the time needed to deliver packets, even if hardware along a path fails or congestion occurs. Current networking technologies follow a reactive approach in which retransmission mechanisms and routing update protocols respond once packet loss occurs. We present a proactive approach that pre-builds and monitors redundant paths to ensure timely delivery with no recovery latency, even in the case of a single-link failure. The paper outlines the approach, describes a new mechanism we call a redundancy controller, and explains the use of SDN to establish and change paths. The paper also reports on a prototype system and gives experimental measurements.