At step 350, the LRM identifies the SRM, which is responsible for the component, via a table lookup or other discovery process, and then calls the SRM using a timeout (in case the SRM process or machine is down), passing information about the failed component.