Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Method for automatic creation and customization of cross-domain symptoms based on resiliency benchmarks

IP.com Disclosure Number: IPCOM000183556D
Original Publication Date: 2009-May-27
Included in the Prior Art Database: 2009-May-27
Document File: 7 page(s) / 62K

Publishing Venue

IBM

Abstract

Disclosed is a method for the capture and population of rules and actions associated to well-known problems executed through resiliency testing. The method will customize cross-domain symptoms for those tests taking into account the entire system environment in which the tests are performed. These cross-domain symptoms can later be used in a production environment to automatically recognize and resolve problems.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 26% of the total text.

Page 1 of 7

Method for automatic creation and customization of cross -domain symptoms based on resiliency benchmarks

Disclosed is a method for the capture and population of rules and actions associated to well-known problems executed through resiliency testing. The method will customize cross-domain symptoms for those tests taking into account the entire system environment in which the tests are performed. These cross-domain symptoms can later be used in a production environment to automatically recognize and resolve problems.

In the IT industry today there is need for automated problem determination and resolution. The autonomic computing symptoms format is a way to specify artifacts used for automated recognition and resolution of problems. The main problem in automated problem determination and resolution is to correctly populate these artifacts, so they can be processed at runtime by correlation engines without human intervention.

There are two main types of symptoms:
- single-domain, where problems are analyzed in the context of a single IT resource (symptoms for a product or a component that does not involve communication/dependency/interaction with other products or components)
- cross-domain, where problems are analyzed in the context of multiple IT resources (typically involving communication or dependency and a specific root-cause symptom can be identified)

Sources of information exist today for population of single-domain symptoms, in the form of product manuals (hard-coded or online), message catalogs, log records, etc. These sources can be instrumented in a semi-automated fashion and converted to symptom catalogs that can be utilized at runtime by the correlation engines.

But the creation of cross-domain symptoms is complex and difficult to achieve. Cross-domain information between multiple products is virtually nonexistent or captures in the heads of system administrators that have accumulated extensive experience with the system over time. Only these people are able to decipher the output of an IT system and take action before an outage occurs. This expertise is often costly to retain, and often lost if key people leaves the support organization.

This disclosure aims to provide a method for automated creation of symptoms in a cross-domain environment, thus automating many of the commonplace tasks performed today by human administrators in an IT support center environment. With the knowledge of the system responses to common disturbances identified and available for systematic data mining, the reliance on human knowledge and the associated risk will be reduced.

The method disclosed here is in fact a very simple but powerful way to create the necessary assets necessary to define symptoms for problem determination and resolution.

1

Page 2 of 7

Figure 1 shows the autonomic computing symptoms format that can be util...