Sheer volumes of data are being generated in extreme-scale distributed scientific applications, and need to be transferred remotely in fast, predictable and reliable way for data storage and analysis purpose. Reserving bandwidth along selected paths in high-performance networks (HPNs) has proved to be an effective way to satisfy the high-demanding performance requirements of such data transfer. However, node and link failures within the HPNs potentially degrade the quality of data transfer. In this paper, we focus on the scheduling of two generic types of bandwidth reservation requests concerning data transfer reliability: (i) to achieve the highest data transfer reliability under a given data transfer deadline, and (ii) to achieve the earliest data transfer completion time while satisfying a given data transfer reliability requirement. Poisson distribution is used to model the node and failures within the HPNs, and two periodic bandwidth reservation algorithms with rigorous optimality proofs are proposed.