Monitoring of distributed systems
AssigneeBritish Telecommunications public limited company
InventorsPaul McKee, Nicholas Davies, Sandra Stin{hacek over (c)}ić-Clarke
Life of this patent
Estimating the term…
Abstract
An environment of sensors and actuators is operated in accordance with predetermined policies, and the effectiveness of individual policies is monitored by comparing inputs from the sensors (550) with intended outcomes from policies controlled by previous inputs from one or more of the same sensors. Each policy includes a check function (554), wherein a sensor, other than the one that triggers the actions, monitors the actions generated by the policy and generates an output (556) indicative of whether a system state is approaching a condition specified by the policy.
Claims12
A network of sensors and actuators for monitoring and controlling an environment in accordance with predetermined policies, the network having a policy decision point for responding to trigger inputs received from the sensors, a data store for maintaining policies for retrieval by the policy decision point, and a policy execution point for controlling actuators in response to commands received from the policy decision point, further comprising a reliability monitoring element which identifies effectiveness of individual policies by identifying an intended outcome from a policy, measuring an input from a sensor following implementation of a policy, and comparing the input from the sensor with the intended outcome, in which each policy includes a check action for assessing the result of that policy's application, using a sensor input to monitor if actions generated by the policy are successful in moving the system towards a system state specified by the policy, in which actions generated by the policy are initiated in response to inputs from one or more first sensors, and the check action is initiated in response to data received from at least one further sensor independent of the one or more first sensors, and in which the check action is initiated after a predetermined period after the actions initiated by the policy, to allow the system to stabilize.
A network of sensors and actuators for monitoring and controlling an environment in accordance with predetermined policies, the network having a policy decision point for responding to trigger inputs received from the sensors, a data store for maintaining policies for retrieval by the policy decision point, and a policy execution point for controlling actuators in response to commands received from the policy decision point, further comprising a reliability monitoring element which identifies effectiveness of individual policies by identifying an intended outcome from a policy, measuring an input from a sensor following implementation of a policy, and comparing the input from the sensor with the intended outcome, in which each policy includes a check action for assessing the result of that policy's application, using a sensor input to monitor if actions generated by the policy are successful in moving the system towards a system state specified by the policy, in which actions generated by the policy are initiated in response to inputs from one or more first sensors, and the check action is initiated in response to data received from at least one further sensor independent of the one or more first sensors, and in which a quality ranking of the policy is generated based on the effectiveness of its action in moving the system towards a system state specified by the policy.
A method of monitoring and controlling an environment using a network of sensors and actuators operating in accordance with predetermined policies, wherein effectiveness of individual policies is monitored by comparing inputs from the sensors with intended outcomes from policies controlled by previous inputs from one or more of the same sensors, in which each policy includes a check action for assessing the result of that policy's application, wherein a sensor monitors actions generated by the policy and generates an output indicative of whether a system state is approaching a condition specified by the policy, and actions generated by the policy are initiated in response to inputs from one or more first sensors, and the check action is initiated in response to data from at least one further sensor independent of the one or more first sensors, and in which the check action is initiated after a predetermined stabilization period after the actions are initiated by the policy.
A method of monitoring and controlling an environment using a network of sensors and actuators operating in accordance with predetermined policies, wherein effectiveness of individual policies is monitored by comparing inputs from the sensors with intended outcomes from policies controlled by previous inputs from one or more of the same sensors, in which each policy includes a check action for assessing the result of that policy's application, wherein a sensor monitors actions generated by the policy and generates an output indicative of whether a system state is approaching a condition specified by the policy, and actions generated by the policy are initiated in response to inputs from one or more first sensors, and the check action is initiated in response to data from at least one further sensor independent of the one or more first sensors, and in which a quality ranking of the policy is generated based on the effectiveness of its action in moving the system towards a system state specified by the policy.
Dependencies are inferred from the claim text.