Browse Prior Art Database

Nonintrusive, Source Code Independent Tracking of Code Flow for a Scenario to diagnose potential abnormal conditions and take corrective action without user interaction

IP.com Disclosure Number: IPCOM000238861D
Publication Date: 2014-Sep-23
Document File: 6 page(s) / 76K

Publishing Venue

The IP.com Prior Art Database

Abstract

When an enterprise application crashes or becomes unresponsive, it leads to a direct loss of revenue. One of the ways to avoid them is to take corrective action when early indications of faults are seen. The description below can be used to diagnose potential abnormal conditions and take corrective action without user interaction.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 32% of the total text.

Page 01 of 6

Nonintrusive, Source Code Independent Tracking of Code Flow for a Scenario to diagnose potential abnormal conditions and take corrective action without user interaction

Even though the existing systems like Emma provide API coverage across the applications to track code flow. These tools are not designed for distributed systems. Though they provide coverage information, the do not give information on dynamic flow of code across a complete lifecycle of execution. Further, most existing code flow tracking tools cannot perform in

providing the code flow events for a scenario when multiple scenarios are executed concurrently.

Other code flow tracking tools monitor and report the execution detail on a per thread basis. Though these tools are primarily designed to tracking the execution of code and report API level metrics, they do not however bring in the notion of a scenario and as such cannot co-relate execution metrics across multiple threads (and or systems) to the scenario (use case or test case) under which they were executed. The system disclosed herein can perform all these operations easily because execution is tracked against a scenario and the system propagates the scenario identity across threads and systems. Therefore data is generated in an inherently co-related manner.

With respect to the fault detection and prevention, the type of faults that the disclosed system can find is more application specific than existing systems. While there are many existing systems that can detect faults in a system, the faults detected are for a thread of execution or for an isolated API. The existing systems do not track the scenario and application context and hence do not consider them when raising faults. So, the faults raised by these systems are isolated and difficult to analyze further. The disclosed system uses the scenario and the application context to identify patterns in the application usage that leads to faults. This makes the type of alerts generated for faults more application specific and can be used to observe broader patterns more easily. For example, patterns such as when an order contains a specific item, the promotion engine performs a memory intensive operation can be easily detected in the disclosed system.

Disclosed is a system to diagnose potential abnormal / fault conditions using Nonintrusive, Source Code independent tracking of Code Flow for a scenario execution. Once the fault condition is observed, the system will take corrective action based on the data that was captured during the tracking of a scenario execution. Scenario execution here can refer to a Testcase or Usecase execution or an end user / customer using the retailer's website.

There are two aspects to the proposed system:

1. Code Flow tracking system: A system that tracks the flow of code for a scenario across its lifecycle of execution.

2. Find potential faults & take corrective action: The Code flow tracking system is monitored live in parallel...