Policy aggregation for computing network health

Abstract

A network health monitoring application computes aggregate reports of network health by combining status results from a group of segments to render a summary health status indicative of the entire group. Individual status values result from policies, which monitor various segments and metrics, or performance parameters (such as byte counts or retransmissions), concerned with providing a particular service. The policies form a hierarchy defining the general or specific nature of the segments included in the policy according to a range of attributes, in which the attributes define a scope of the segments included in the policy group. Multiple metrics gathered on each segment allow filtering by polices based on the metric, rather than the attribute (hierarchy level), thus lending a hierarchy filtering based on 'vertical' slices of the hierarchy, allowing the user flexibility to filter on specific attribute values while still computing aggregate health at a particular attribute level.