Given a requirement for reliable in-app/cluster message handoff that must survive restarts, identify a persistent (and where needed transactional) VM queue rather than a transient VM queue.