EXFO has been granted a patent for a computer-implemented method that analyzes data records to identify hardware or software components causing communication problems in a network. The method involves determining the degree of association between signatures in the data records and communication problems, grouping signatures into equivalence classes, generating a dependency graph, and selecting components related to the network that are likely involved in the communication problems. GlobalData’s report on EXFO gives a 360-degree view of the company including its patenting strategy. Buy the report here.
According to GlobalData’s company profile on EXFO, AI-assisted repair estimation was a key innovation area identified from patents. EXFO's grant share as of September 2023 was 66%. Grant share is based on the ratio of number of grants to total number of patents.
Patent granted for identifying incompatible hardware or software components
A recently granted patent (Publication Number: US11736339B2) describes a computer-implemented method for analyzing communication sessions in a network to identify incompatible hardware or software components. The method involves obtaining a set of data records that characterize the operation of communication sessions in the network. These data records contain tuples with signatures representing features and values related to the network components involved in the sessions.
The method then determines a 2-signature tuple, composed of a first signature and a second signature, present in the data records. It calculates a first gain, which represents the overall relative inefficiency of the communication sessions involving the 2-signature tuple compared to the relative inefficiencies of the sessions involving the first or second signature individually. If the first gain exceeds a predetermined threshold, the method filters the communication sessions involving the 2-signature tuple to create a subset of sessions involving 1-signatures. The size of this subset must exceed a second predetermined threshold.
Next, the method calculates a second gain, representing the overall relative inefficiency of the communication sessions involving the 2-signature tuple compared to the relative inefficiencies of the sessions involving the 1-signatures in the subset. If the second gain exceeds the first predetermined threshold, the method identifies the features and values represented by the first and second signatures as units of the hardware or software components that are incompatible.
The patent also describes additional operations, such as identifying equivalence classes of incompatible 1-signatures, generating dependency graphs between these signatures, and removing nodes from the graphs based on compatibility changes.
Overall, this patented method provides a computer-based approach to analyze communication sessions in a network and identify incompatible hardware or software components. By calculating gains and thresholds, the method can efficiently filter and identify problematic components, allowing for targeted troubleshooting and optimization efforts.