Failure recovery is a critical performance issue in large-scale Ethernet networks. The standard mechanisms of link-failure recovery in spanning tree protocol (STP) and rapid STP may cause interruptions of data transmission because of long recovery time. Several previous approaches construct multiple alternative trees to avoid such interruptions caused by single or multiple failed links based on multiple STP. However, the prototypical protocol of standard root election would result in the same root switch for different spanning trees. Hence, the alternative trees may become invalid if a failed link appears in more than 1 spanning tree. Our scheme, failure-adaptive spanning trees (FAST), improves the resilience of Ethernet with multiple spanning trees. It avoids simultaneous faults of multiple spanning trees by incorporating a new approach to electing an appropriate root switch for each spanning tree. Failure-adaptive spanning trees also use active links of alternative spanning trees for data transmission. The paths of different spanning trees can thus act as shortcuts to shorten frame paths. The experimental results demonstrate that FAST can sustain network throughput under multiple link failures. The FAST also balances both link and switch loads. KEYWORDS ethernet, failure recovery, load balancing, resilience, spanning tree 1 Int J Commun Syst. 2017;30:e3291.wileyonlinelibrary.com/journal/dac