Abstract: The connectivity information provided by a variety of inference engines is integrated to provide a set of inferred links within a network. A consolidation is performed among inference engines that operate at a base level of connectivity detail to create a model of the network at this base level. The connectivity information provided by inference engines at each subsequent higher level of connectivity abstraction is then overlaid on the base level connectivity. By separately consolidating the connectivity information at each level of abstraction, the rules for dealing with conflicts can be simplified and/or better focused to resolve the conflict. By assuming that the more detailed lower level information is likely to be more accurate, rules can be developed to modify the connectivity models produced by the higher level techniques to conform to the lower level connectivity details while still maintaining the integrity of the higher level connectivity models.

Abstract: The physical connection corresponding to IP tunnels in a network are found by tracing through the device configuration and routing tables at the routers in the network to determine the outbound interface associated with each tunnel endpoint, and then inferring a likely return interface associated with the opposite tunnel endpoint. Having determined the physical devices at the source and destination of each tunnel, the physical path between these source and destination devices is traced from the source toward the destination until the path is terminated at the destination device, or at an interface to an external network. If the path ends at an external network, the path is traced from the destination device toward the source device until a corresponding interface to the external network is reached.

Abstract: The connectivity information provided by a variety of inference engines is integrated to provide a set of inferred links within a network. A consolidation is performed among inference engines that operate at a base level of connectivity detail to create a model of the network at this base level. The connectivity information provided by inference engines at each subsequent higher level of connectivity abstraction is then overlaid on the base level connectivity. By separately consolidating the connectivity information at each level of abstraction, the rules for dealing with conflicts can be simplified and/or better focused to resolve the conflict. By assuming that the more detailed lower level information is likely to be more accurate, rules can be developed to modify the connectivity models produced by the higher level techniques to conform to the lower level connectivity details while still maintaining the integrity of the higher level connectivity models.

Abstract: A network is partitioned into a set of independent partitions, and the topology of each partition is determined, then merged to form a topology of the entire network. Preferably, the partitioning is hierarchical, wherein the network is partitioned to form individual VLAN partitions, and each of the VLAN partitions is further partitioned based on the nodes that are simply connected to each port of one or more selected root switches within the VLAN partition. Simple connections to each port are efficiently determined based on an aggregate address forwarding table associated with each node. Ancillary information, such as spanning tree or CDP data, may be used to facilitate efficient partitioning and/or to validate inferences that are made with incomplete information.

Abstract: A network analysis system invokes an application specific, or source-destination specific, path discovery process. The application specific path discovery process determines the path(s) used by the application, collects performance data from the nodes along the path, and communicates this performance data to the network analysis system for subsequent performance analysis. The system may also maintain a database of prior network configurations to facilitate the identification of nodes that are off the path that may affect the current performance of the application. The system may also be specifically controlled so as to identify the path between any pair of specified nodes, and to optionally collect performance data associated with the path.

Abstract: A network analysis system invokes an application specific, or source-destination specific, path discovery process. The application specific path discovery process determines the path(s) used by the application, collects performance data from the nodes along the path, and communicates this performance data to the network analysis system for subsequent performance analysis. The system may also maintain a database of prior network configurations to facilitate the identification of nodes that are off the path that may affect the current performance of the application. The system may also be specifically controlled so as to identify the path between any pair of specified nodes, and to optionally collect performance data associated with the path.

Abstract: The physical connection corresponding to IP tunnels in a network are found by tracing through the device configuration and routing tables at the routers in the network to determine the outbound interface associated with each tunnel endpoint, and then inferring a likely return interface associated with the opposite tunnel endpoint. Depending upon the particular configurations, a variety of tests can be applied to validate the inference. Patricia trees are preferably used to store and process the configuration data for efficient tracing through the routing tables at each router.

Abstract: A method and/or system is configured to improve the results of an auto-detection of network devices based on the causes of detection failures in preceding runs of the auto-detection process. As each device that is believed to be in the network is found to be undiscovered, the identification of the device and information regarding the cause(s) of non-discovery are stored. Prior to the next auto-detection run, one or more of the discovery parameters are modified, based on the causes associated with the undiscovered devices. The extent to which the discovery parameters are modified is preferably based on the apparent stability of the network, or upon the detection of changes to the network.

Abstract: The connectivity information provided by a variety of inference engines is integrated to provide a set of inferred links within a network. A consolidation is performed among inference engines that operate at a base level of connectivity detail to create a model of the network at this base level. The connectivity information provided by inference engines at each subsequent higher level of connectivity abstraction is then overlaid on the base level connectivity. By separately consolidating the connectivity information at each level of abstraction, the rules for dealing with conflicts can be simplified and/or better focused to resolve the conflict. By assuming that the more detailed lower level information is likely to be more accurate, rules can be developed to modify the connectivity models produced by the higher level techniques to conform to the lower level connectivity details while still maintaining the integrity of the higher level connectivity models.

Abstract: The physical connection corresponding to IP tunnels in a network are found by tracing through the device configuration and routing tables at the routers in the network to determine the outbound interface associated with each tunnel endpoint, and then inferring a likely return interface associated with the opposite tunnel endpoint. Having determined the physical devices at the source and destination of each tunnel, the physical path between these source and destination devices is traced from the source toward the destination until the path is terminated at the destination device, or at an interface to an external network. If the path ends at an external network, the path is traced from the destination device toward the source device until a corresponding interface to the external network is reached.

Abstract: A network is partitioned into a set of independent partitions, and the topology of each partition is determined, then merged to form a topology of the entire network. Preferably, the partitioning is hierarchical, wherein the network is partitioned to form individual VLAN partitions, and each of the VLAN partitions is further partitioned based on the nodes that are simply connected to each port of one or more selected root switches within the VLAN partition. Simple connections to each port are efficiently determined based on an aggregate address forwarding table associated with each node. Ancillary information, such as spanning tree or CDP data, may be used to facilitate efficient partitioning and/or to validate inferences that are made with incomplete information.