Abstract: A method of determining computer network topologies that dramatically reduces the computational complexity and greatly increases the accuracy of connection determination. The method involves classifying ports as either up or down. A source address table is compiled for each port of each data-relay device and each port is classified as either up or down. Up ports connect to other data-relay devices which report source address tables while down ports do not. After the classification, each source address in each up port table is replaced by the source address of the data-relay devices containing the down port whose table contains that source address. The tables of pairs of up ports are compared by intersection and minimal intersection defines the most probable connection for each up port. A variety of methods are used to remove invalid source addresses and the addresses of devices that have moved during the collection of the source address tables.
Abstract: A method of analyzing a communication network comprising determining a mean drop rate in a device x by polling each device from a network management computer (NMC) which is in communication with the network, and processing signals in the NMC to determine a drop rate D(x), in accordance with:
D(x)=((L+(x)−L−(x))/2,
and
L(x)=1−A(x)
where
A(x): the fraction of poll requests from the NMC to device x for which the NMC receives replies (measured over the last M sampling periods), (wherein x must not be broken),
D(x): the mean frame drop rate in device x,
L(c): NMC's perception of the loss rate to device x and back,
L−(x): the NMC's perception of the mean value of L(z) for all devices z connected to device x, closer to the NMC than device x and which are not broken, and
L+(x): the NMC's perception of the mean value of L(z) for all devices z connected to device x, further away from the NMC than device x and which are not broken.
Abstract: A method of detecting outliers measured during progression of an activity of an entity from one point to another point, comprising measuring activity at a point in a first dimension, measuring the same activity at the same point in at least a second dimension referenced to the same time as measuring the activity in the first dimension, and rejecting outliers which have values outside a maximum expected difference between the activity measured in the first and second dimensions.
Abstract: A method for automatically aggregating objects and determining a hierarchical organization of these objects and aggregates by exploiting known or computed priorities and correlations.
Abstract: A method of analyzing a communication network that determines a mean drop rate in a device x by polling each device from a network management computer (NMC) which is in communication with the network, and processing signals in the NMC to determine a drop rate D(x), in accordance with:D(x)=((L+(x)-L-(x))/2,and L(x)=1-A(x)whereA(x): the fraction of poll requests from the NMC to device x for which the NMC receives replies (measured over the last M sampling periods), (wherein x must not be broken),D(x): the mean frame drop rate in device x,L(c): NMC's perception of the loss rate to device x and back,L-(x): the NMC's perception of the mean value of L(z) for all devices z connected to device x, closer to the NMC than device x and which are not broken, andL+(x): the NMC's perception of the mean value of L(z) for all devices z connected to device x, further away from the NMC than device x and which are not broken.
Abstract: A hybrid method is described that allows for the combination of both source-compiled and source-interpreted code in the execution of computer programs. While traditional practice has dictated otherwise, code does not need to be either solely interpreted or solely compiled.