Internet of Things gateways, edge authentication servers, and legacy systems face a serious threat: offline password guessing after a credential-database disclosure. Memory-hard key derivation functions (KDFs) are a common defense, but their computational and memory costs can be prohibitive for resource-constrained devices.

The paper introduces ARMOR, a retrofit designed to amplify the adversary's cost in such environments. Because it is schema-preserving, ARMOR can be applied without altering the existing password storage schema, making it practical for systems that cannot easily migrate to heavier KDFs.

By focusing on the gap between security needs and resource limits, ARMOR offers a way to harden password storage without requiring the full overhead of memory-hard functions. The work specifically addresses the constraints of IoT and edge deployments, where traditional defenses may be too expensive to run. Further details on the mechanism and evaluation are in the full paper.