Resilient routing in transport networks requires the calculation of active paths (APs) and backup paths (BPs) to be used in the event of failures in the corresponding APs. Often it is not possible to ensure full end-to-end protection. Hence an AP is only an admissible solution if a BP can be calculated such that it shares a minimal number of common nodes with the AP. In the absence of node disjoint path pairs, the existence of parallel edges has to be taken into consideration, because parallel edges could ensure link protection between two adjacent nodes. The potential advantages of using multicriteria models in various routing contexts has been pointed out by several authors. In this paper we present a bicriteria optimization model for calculating APs with maximally disjoint BPs. A network performance study using a real transport network, enabling a comparison of the model results with those of the associated single criterion models, is carried out. The two considered metrics are hop count and load cost. Having calculated an AP, the best BP was subsequently obtained (using the same metric). The network was represented using a directed multi-graph, because the existence of parallel arcs between end nodes had to be explicitly addressed and connection requests for services with asymmetrical bandwidth requirements were also considered. A major objective of this work was the improvement of protected path selection, from a network management point of view, where an answer to clients requests for new connections, in transport networks, has to be answered in less than a second.