Browse Prior Art Database

A Method and System for Issue Insight Analysis

IP.com Disclosure Number: IPCOM000244873D
Publication Date: 2016-Jan-25
Document File: 5 page(s) / 276K

Publishing Venue

The IP.com Prior Art Database

Abstract

Many industries (e.g., banking, telecom) heavily rely on IT systems to provide business services to huge users. However, lots of IT issues would happen in enterprise IT environments, which not only largely degrades users' satisfactory but also affects enterprises' revenue. To help enterprise accounts to better remediate/prevent IT issues during their daily IT maintenance, we propose a general-purpose issue insight analysis framework for enterprise accounts. In our framework, starting with issues stored in issue tracking systems, we incrementally extract and polish the issue models based on cross-reference with other knowledge bases stored in other enterprise management systems (e.g., Configuration Management Database (CMDB), Source Code Management System (SCM), Human Resource management system (HRMS)). The generated/polished issue models provide good foundations for the extraction processes of various types of issue insights.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 01 of 5

A Method and System for Issue Insight Analysis
A. Problem and Motivation

As shown in Fig. 1, an issue is usually associated with multiple categories of important entities (e.g., symptoms, prompting activities, impacts, root causes, remediation actions, related experts/developers, affected servers/nodes, systems, components/features). Once we get such ontology models of issues, it would be applicable to generate various types of insights on top of them (e.g., what symptoms indicate a specific problem, what systems require corrective actions).

Within enterprises, issue tracking systems provide central places to create/store their IT maintenance issues. The issues stored in these systems are good starting points toextract issue insights. However, there are some obvious limitations when using these issue items exported from issue tracking systems to extract insights:

1. Most of the entities are not explicitly provided as stand-alone attributes by each issue. They may be embedded within the long description info of each issue, but require innovative techniques to extract them out smartly.


2. Many embedded entities are described with abstract/elusive numbers, which poses barriers for the automatic

1

Figure 1. Issue Ontology Model


Page 02 of 5

understanding.

3. Many entities cannot be extracted due to incomplete descriptions of issues.

To address these limitations, we propose a general-purpose issue insight analysis framework for enterprise accounts. In our framework, starting with issues stored in issue tracking systems, we incrementally extracted and polish the issue models based on cross-reference with other knowledge bases stored in other enterprise management systems (e.g., Configuration Managemetn database (CMDB), source code management system (SCM), human resource management system (HRMS)). Powered by the insights extracted with our analysis framework, enterprise accounts can remediate/prevent IT issues more effectively and efficiently during their daily IT maintenance.


B. Approach

2


Page 03 of 5

Figure 2. Issue Insight Analysis Framework

As shown in Fig. 2, our frameworks includes five major steps:
a.
Issue Extraction and Normalization.In this step, we first apply different adapters to extract original issue items from

different issue tracking systems or files (e.g., databases, excel files, txt files), then apply corresponding normalizers to transform these original issues as normalized issue items with standard attributes (e.g., MUST fields such as ID, creation timestamps, creator, title, description and etc., and OPTIONAL fields such server/node name, system name, component name, source, priority, handlers, causes, actions, future suggestions and etc.). With the normalized issue items, we identify redundant issues based on a serious of heuristics, and then remove redundant issues.


b. Issue Concept Extraction.In this step, we use entity resolution techniques to extract important types of entities (e.g.,

3


Page 04 of 5

associated...