A hierarchical multicast monitoring scheme

Deployment of multicast routing services in corporate networks and Internet Service Providers is still tentative. Among other problems, there is a lack of monitoring and management tools and systems. Previous work in multicast management has failed to address the scalability problem present in multicast fault isolation and reporting. We propose a hierarchical, passive monitoring scheme, HPMM, that relies on a series of pre-deployed, self-organized monitoring daemons. With HPMM, fault message aggregation and local fault detection and isolation is more efficient than previous approaches. HPMM satisfies a number of design goals: scalable reporting; fault isolation; no dependencies on multicast routing for reporting; and no modifications to existing routing or diagnosis protocols are required. The tradeoff of using HPMM is it leverages a large number of software daemons deployed through out a local domain. We compare the signalling overhead of HPMM and previous work with a simulation.