# US 2019/0171622 A1 · System event search based on heterogeneous logs

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2019/0171622 A1
- Document ID: US20190171622A1
- Status: Published application
- Kind: A1
- Application number: 16/203,008
- Filed: 2018-11-28
- Published: 2019-06-06
- Earliest claimed: 2017-12-04
- Assignees: NEC Laboratories America, Inc.
- Applicants: NEC Laboratories America, Inc.
- Inventors: Bo Zong; Jianwu Xu; Haifeng Chen
- Web page: https://patentgrep.com/patent/US20190171622A1

## CPC

- G06F 16/2455: Query execution
- G06F 11/08: Error detection or correction by redundancy in data representation, e.g. by using checking codes
- G06F 16/152: {using file content signatures, e.g. hash values}
- G06F 16/1734: {Details of monitoring file system events, e.g. by the use of hooks, filter drivers, logs}
- G06F 16/24542: {Plan optimisation}
- G06F 16/2465: {Query processing support for facilitating data mining operations in structured databases}
- G06F 16/9038: Presentation of query results
- G06F 18/285: {Selection of pattern recognition techniques, e.g. of classifiers in a multi-classifier system}
- G06F 2216/03: Data mining
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

Systems and methods for system event searching based on heterogeneous logs are provided. A system can include a processor device operatively coupled to a memory device wherein the processor device is configured to mine a variety of log patterns from various of heterogeneous logs to obtain known-event log patterns and unknown-event log patterns, as well as to build a weighted vector representation of the log patterns. The processor device is also configured to evaluate a similarity between the vector representation of the unknown-event and known-event log patterns, identify a known event that is most similar to an unknown event to troubleshoot system faults based on past actions for similar events to improve an operation of a computer system.

## Claims

### Claim 1

A method for system event searching based on heterogeneous logs, the method comprising:
mining, by the processor device based on instructions stored on a storage device coupled to the processor device, a plurality of log patterns from a first plurality of heterogeneous logs and a second plurality of heterogeneous logs to obtain a set of known-event log patterns for each known event and a set of unknown-event log patterns for each unknown event;
building, by the processor device, a weighted vector representation of each set of known-event log patterns and of each set of unknown-event log patterns;
evaluating, by the processor device, a similarity of the vector representation of each set of unknown-event log patterns to the vector representation of each set of known-event log patterns; and
identifying, by the processor device, a known event that is most similar to an unknown event based on the evaluated similarity to troubleshoot system faults based on past actions for similar events to improve an operation of a computer system.

### Claim 2 (depends on 1)

The method as recited in claim 1, wherein mining each of the plurality of heterogeneous logs of known events and the plurality of heterogeneous logs of unknown events includes mining at least one of text-based logs and time-series-based logs.

### Claim 3 (depends on 1)

The method as recited in claim 1, wherein the plurality of log patterns includes different categories of patterns.

### Claim 4 (depends on 1)

The method as recited in claim 1, wherein building the weighted vector representation includes denoting a specific log pattern for each dimension of the vector representation and setting the vector representation's dimensionality being defined by a number of log patterns within the set of log patterns the vector representation represents.

### Claim 5 (depends on 1)

The method as recited in claim 1, wherein building the weighted vector representation includes a value of the vector representation in each dimension being defined by the number of times a specific log pattern appears within the set of log patterns the vector representation represents.

### Claim 6 (depends on 1)

The method as recited in claim 1, wherein mining is performed via Next Generation Log Analytics (NGLA).

### Claim 7 (depends on 6)

The method as recited in claim 6, wherein building a vector representation includes translating an obtained set of log patterns into a different format that is feasible to be input into a System Event Search (SES) tool by decoupling NGLA pattern mining logic from the logic of the SES searching algorithm.

### Claim 8 (depends on 6)

The method as recited in claim 6, wherein building a vector representation includes generating a vector representation from a translated set of log patterns whereby a vector representation denotes an event's signature.

### Claim 9 (depends on 6)

The method as recited in claim 6, further including storing event signatures in a signature database.

### Claim 10 (depends on 6)

The method as recited in claim 6, wherein evaluating the similarity includes matching signatures of unknown events to signatures of unknown events.

### Claim 11 (depends on 6)

The method as recited in claim 6, further including saving matched events according to a file format schema and displaying troubleshooting options via a display device.

### Claim 12

A computer system for searching system events based on heterogeneous logs, comprising:
a processor device operatively coupled to a memory device, the processor device being configured to:
mine a plurality of log patterns from a first plurality of heterogeneous logs and a second plurality of heterogeneous logs to obtain a set of known-event log patterns for each known event and a set of unknown-event log patterns for each unknown event;
build a weighted vector representation of each set of known-event log patterns and of each set of unknown-event log patterns;
evaluate a similarity of the vector representation of each set of unknown-event log patterns to the vector representation of each set of known-event log patterns; and
identify a known event that is most similar to an unknown event based on the evaluated similarity to troubleshoot system faults based on past actions for similar events to improve an operation of a computer system.

### Claim 13 (depends on 12)

The system as recited in claim 12, wherein each of the plurality of heterogeneous logs of known events and the plurality of heterogeneous logs of unknown events include at least one of text-based logs and time-series-based logs.

### Claim 14 (depends on 12)

The system as recited in claim 12, wherein the plurality of log patterns includes different categories of patterns.

### Claim 15 (depends on 12)

The system as recited in claim 12, wherein each dimension of the vector representation denotes a specific log pattern and wherein the vector representation's dimensionality is defined by a number of log patterns within the set of log patterns the vector representation represents.

### Claim 16 (depends on 12)

The system as recited in claim 12, wherein a value of the vector representation in each dimension is defined by the number of times a specific log pattern appears within the set of log patterns the vector representation represents.

### Claim 17

A computer program product for searching system events based on heterogeneous logs, the computer program product comprising a non-transitory computer readable storage medium having program instructions embodied therewith, the program instructions executable by a computing device to cause the computing device to perform the method comprising:
mining, by the processor device based on instructions stored on a storage device coupled to the processor device, a plurality of log patterns from a first plurality of heterogeneous logs and a second plurality of heterogeneous logs to obtain a set of known-event log patterns for each known event and a set of unknown-event log patterns for each unknown event;
building, by the processor device, a weighted vector representation of each set of known-event log patterns and of each set of unknown-event log patterns;
evaluating, by the processor device, a similarity of the vector representation of each set of unknown-event log patterns to the vector representation of each set of known-event log patterns; and
identifying, by the processor device, a known event that is most similar to an unknown event based on the evaluated similarity to troubleshoot system faults based on past actions for similar events to improve an operation of a computer system.

### Claim 18 (depends on 17)

The product as recited in claim 17, wherein mining is performed via Next Generation Log Analytics (NGLA).

### Claim 19 (depends on 17)

The product as recited in claim 17, wherein building a vector representation includes translating an obtained set of log patterns into a different format that is feasible to be input into a System Event Search (SES) tool by decoupling NGLA pattern mining logic from the logic of the SES searching algorithm.

### Claim 20 (depends on 17)

The product as recited in claim 17, wherein building a vector representation includes generating a vector representation from a translated set of log patterns whereby a vector representation denotes an event's signature.
