# US 2015/0106794 A1 · Transparent performance inference of whole software layers and context-sensitive performance debugging

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

- Number: US 2015/0106794 A1
- Document ID: US20150106794A1
- Status: Published application
- Kind: A1
- Application number: 14/512,653
- Filed: 2014-10-13
- Published: 2015-04-16
- Earliest claimed: 2013-10-14
- Assignees: NEC Laboratories America, Inc.
- Applicants: NEC Laboratories America, Inc.
- Inventors: Junghwan Rhee; Hui Zhang; Nipun Arora; Guofei Jiang; Chung Hwan Kim
- Web page: https://patentgrep.com/patent/US20150106794A1

## CPC

- G06F 11/3636: {by tracing the execution of the program}
- G06F 11/3419: {by assessing time}
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

Methods and systems for performance inference include inferring an internal application status based on a unified call stack trace that includes both user and kernel information by inferring user function instances. A calling context encoding is generated that includes information regarding function calling paths. Application performance is analyzed based on the encoded calling contexts. The analysis includes performing a top-down latency breakdown and ranking calling contexts according to how costly each function calling path is.

## Claims

### Claim 1

A method for performance inference, comprising:
inferring an internal application status based on a unified call stack trace that includes both user and kernel information by inferring user function instances;
generating a calling context encoding that includes information regarding function calling paths; and
analyzing application performance based on the encoded calling contexts, comprising:
performing a top-down latency breakdown; and
ranking calling contexts according to how costly each function calling path is.

### Claim 2 (depends on 1)

The method of claim 1, wherein inferring an internal application status comprises inferring latencies for individual function calls.

### Claim 3 (depends on 2)

The method of claim 2, further comprising annotating the calling context encoding with inferred performance information.

### Claim 4 (depends on 1)

The method of claim 1, wherein performing a top-down latency breakdown comprises:
estimating a total running time of each function in a stack trace;
excluding running times attributable to later-starting, concurrently running functions to determine a latency attributable to each function.

### Claim 5 (depends on 4)

The method of claim 4, wherein estimating a total running time comprises finding a difference between a first stack trace in which a function appears and a first stack trace in which the function no longer appears.

### Claim 6 (depends on 4)

The method of claim 4, wherein estimating a total running time comprises finding a difference between a first stack trace in which a function appears and a last uninterrupted stack trace in which the function appears.

### Claim 7 (depends on 1)

The method of claim 1, wherein ranking calling contexts comprises ranking according to total cost for each function calling path.

### Claim 8 (depends on 1)

The method of claim 1, wherein ranking calling contexts comprises ranking according to a difference in latency between two input workloads.

### Claim 9

A system for performance inference, comprising:
an inference module comprising a processor configured to infer an internal application status based on a unified call stack trace that includes both user and kernel information by inferring user function instances and to generate a calling context encoding that includes information regarding function calling paths; and
a performance analysis module configured to analyze application performance based on the encoded calling contexts by performing a top-down latency breakdown and ranking calling contexts according to how costly each function calling path is.

### Claim 10 (depends on 9)

The system of claim 9, wherein the inference module is further configured to infer latencies for individual function calls.

### Claim 11 (depends on 10)

The system of claim 10, further comprising annotating the calling context encoding with inferred performance information.

### Claim 12 (depends on 9)

The system of claim 9, wherein performing a top-down latency breakdown comprises:
estimating a total running time of each function in a stack trace;
excluding running times attributable to later-starting, concurrently running functions to determine a latency attributable to each function.

### Claim 13 (depends on 12)

The system of claim 12, wherein estimating a total running time comprises finding a difference between a first stack trace in which a function appears and a first stack trace in which the function no longer appears.

### Claim 14 (depends on 12)

The system of claim 12, wherein estimating a total running time comprises finding a difference between a first stack trace in which a function appears and a last uninterrupted stack trace in which the function appears.

### Claim 15 (depends on 9)

The system of claim 9, wherein ranking calling contexts comprises ranking according to total cost for each function calling path.

### Claim 16 (depends on 9)

The system of claim 9, wherein ranking calling contexts comprises ranking according to a difference in latency between two input workloads.
