Open research directions for the evidence-governed OT deployment framework

Establish convergence guarantees for the repair cycle, develop federated evidence pooling, derive online clause-learning methods for behavior-only adaptation, construct hook-independent attestation mechanisms, and create a shared benchmark for industrial operational-technology networks.

Background

The paper presents an evidence-governed deployment framework for networked workloads and self-adapting agents operating on shared industrial operational-technology infrastructure. Its workflow uses a versioned network-intent artifact, offline evidence qualification, readiness assessment, cross-workload repair, and passive drift monitoring. The framework is intended to distinguish demonstrated readiness, genuine violations, and undecidable cases caused by missing evidence.

The authors identify several unresolved extensions. They do not provide convergence guarantees for repeated or iterative repair operations; federated evidence pooling is not addressed; the framework lacks online learning of requirements from behavior-only adaptations; it does not provide attestation independent of local hooks; and the field lacks a shared OT benchmark for evaluating such deployment-readiness systems. These topics are explicitly deferred to future work.