Universal network language model with API interface for cross-platform support and log summarization

Patent metadata

Field	Value
Publication number	US-12574284-B2
Application number	US-202418771762-A
Country	US
Kind code	B2
Filing date	Jul 12, 2024
Priority date	Jul 12, 2024
Publication date	Mar 10, 2026
Grant date	Mar 10, 2026

Abstract

A computing system for cross-platform support and log summarization includes a language model, APIs, processors, and memories to generate queries and troubleshooting steps from log files and IR system results. A method involves obtaining log files, generating IR system queries, and creating troubleshooting steps using a language model. A non-transitory computer readable medium contains instructions for obtaining log files, generating queries, and creating troubleshooting steps based on log files and IR results.

First claim

What is claimed:

1 . A computing system for cross-platform support and log summarization comprising:
a language model configured to generate information retrieval (IR) system queries and recommended troubleshooting steps for a plurality of networks and respective network configurations; one or more application programming interfaces (API) communicatively interfacing the language model and an IR system and configured to obtain responsive files from the IR system using one or more IR system queries from the language model; one or more processors; and one or more memories having stored thereon computer executable instructions that, when executed by the one or more processors, cause the computing system to: obtain, by an API and from a plurality of devices, a plurality of log files associated with a network and a respective network configuration; input, to the language model, the plurality of log files; generate, by the language model, one or more recommended IR system queries for each log file; input, to the language model and by an API, IR system results associated with the IR system queries for each log file; generate, by the language model, one or more recommended troubleshooting steps based on the plurality of log files and the IR system results; generate, by a computer vision API, a summary for each of one or more network topology diagrams associated with the respective network configuration; and input, to the language model and with the plurality of log files, the summary.

2 . The computing system of claim 1 , wherein the plurality of log files are obtained in response to user input and wherein the user input includes one or more questions about the network.

3 . The computing system of claim 2 , wherein the user input includes the one or more network topology diagrams.

4 . The computing system of claim 1 , further comprising instructions that, when executed by the one or more processors, cause the computing system to:
input, to the language model and with the plurality of log files, additional network data including active device states and network interface configurations.

5 . The computing system of claim 1 , wherein the one or more API obtains the IR system results from: (i) a search engine, or (ii) a database of the computing system utilizing query-driven retrieval.

6 . A computer-implemented method for cross-platform support and log summarization, comprising:
obtaining, by an application programming interface (API) communicatively interfacing a language model configured to generate information retrieval (IR) system queries and recommended troubleshooting steps for a plurality of networks and respective network configurations, and an IR system,
wherein the API is configured to obtain responsive files from the IR system using one or more IR system queries from the language model, and from a plurality of devices, a plurality of log files associated with a network and a respective network configuration;
inputting, to the language model, the plurality of log files; generating, by the language model, one or more recommended IR system queries for each log file; inputting, to the language model and by an API, IR system results associated with one or more IR system queries for each log file; generating, by the language model, one or more recommended troubleshooting steps based on the plurality of log files and the IR system results; generating, by a computer vision API, a summary for each of one or more network topology diagrams associated with the respective network configuration; and inputting, to the language model and with the plurality of log files, the summary.

7 . The computer-implemented method of claim 6 , wherein the plurality of log files are obtained in response to user input and wherein the user input includes one or more questions about the network.

8 . The computer-implemented method of claim 6 , further comprising:
inputting, to the language model and with the plurality of log files, additional network data including active device states and network interface configurations.

9 . The computer-implemented method of claim 6 , wherein the one or more API obtains the IR system results from: (i) a search engine, or (ii) a database utilizing query-driven retrieval.

10 . A non-transitory computer readable medium containing program instructions that when executed by one or more processors, cause a computer to:
obtain, by an application programming interface (API) communicatively interfacing a language model configured to generate information retrieval (IR) system queries and recommended troubleshooting steps for a plurality of networks and respective network configurations, and an IR system,
wherein the API is configured to obtain responsive files from the IR system using one or more IR system queries from the language model, and from a plurality of devices, a plurality of log files associated with a network and a respective network configuration;
input, to the language model, the plurality of log files; generate, by the language model, one or more recommended IR system queries for each log file; input, to the language model and by an API, IR system results associated with one or more IR system queries for each log file; generate, by the language model, one or more recommended troubleshooting steps based on the plurality of log files and the IR system results; generate, by a computer vision API, a summary for each of one or more network topology diagrams associated with the respective network configuration; and input, to the language model and with the plurality of log files, the summary.

11 . The non-transitory computer readable medium of claim 10 , wherein the plurality of log files are obtained in response to user input and wherein the user input includes one or more questions about the network.

12 . The non-transitory computer readable medium of claim 11 , wherein the user input includes the one or more network topology diagrams.

13 . The non-transitory computer readable medium of claim 10 , including further instructions that, when executed by the one or more processors, cause a computer to:
input, to the language model and with each respective log, additional network data including active device states and network interface configurations.

Assignees

• Cdw Llc

Inventors

• Gruber Carl

Classifications

• G06F16/345

  Summarisation for human users · CPC title

• H04L41/16

  using machine learning or artificial intelligence · CPC title

• H04L41/149

  for prediction of maintenance · CPC title

• H04L41/069 Primary

  using logs of notifications; Post-processing of notifications · CPC title

• H04L41/12 Primary

  Discovery or management of network topologies · CPC title