Naming attribute standardization

Discovered data can be imported into Integration Composer from different data sources and with multiple discovery engines. However, by using various sources and tools, the same record attributes can be structured in different ways. As a result, duplicate entries can be created in the Maximo® database. In Integration Composer, rules are defined that standardize the structure of naming attribute data. The standardization is done before records are reconciled and made available to Control Desk.

When data is imported by Integration Composer, additional cleansing rules are applied to the naming attribute values for each record, standardizing their structure. The record is then registered with the Naming and Reconciliation Service (NRS) and assigned an integration identifier. Finally, records are inserted into the Maximo database and reconciled as deployed assets and actual CIs, based on their assigned integration identifiers.

If you loaded deployed asset data, with a version of Integration Composer that is earlier than 7.5, you can have duplicate records:

For version 7.5, you must reload the data through Integration Composer, so that the data can be standardized and matched appropriately.
For version 7.5.1 and higher, a new Asset and CI Cleansing and Reconciliation tool is available that takes advantage of further enhancements to NRS. The tool cleanses your existing assets and CIs and generates standardized integration identifiers, which are stored in your Maximo database.

Naming attribute cleansing rules

The cleansing rules are defined in a customizable .xml file, which is held on the Integration Composer server in this location:

<Integration Composer install directory>\data\properties\nrs\cleansingRules.xml

This file is also held on the Control Desk server:

<Control Desk install directory>\maximo\applications\maximo\properties\sccd\

The file contains rules for data imported with these discovery engines but can be modified to define standardization rules for your own engines:

IBM® Tivoli® Asset Discovery for Distributed
IBM Tivoli Application Dependency Discovery Manager (TADDM)

If changes are made to the cleansing rule file, the file must be copied to both the Integration Composer and Control Desk servers. For the Control Desk server, you must also rebuild and redeploy the maximo.ear.

Rule file structure

The cleansing rules file uses the following structure:

<Rule major="ComputerSystem" minor="SerialNumber">
	     <Regex pattern="&quot;" format="" />
 	     <Regex pattern="_" format=" " />
	     <Regex pattern="^VMware-" format="" />
	     <Trim />
	     <UpperCase />
</Rule>

Where the element <Rule> defines a single naming attribute. There are multiple <Rule> elements in the file. Each <Rule> consists of the following elements and attributes and are run in the sequence in which they are defined:

major identifies the Common Data Model class, containing the naming attribute to be cleansed. Currently, ComputerSystem is the only supported value.
minor identifies the name of the naming attribute to be cleansed with the following rules.
<Regex> statements define regular expression search-and-replace functions, with the replaceAll() method of the Java String class. It has two attributes:
- pattern specifies the value to search for in the minor naming attribute.
- format specifies the substitution value that is used to replace the pattern.
In the example:
- The first, pattern=""" format="", removes double quotation marks.
- The second, pattern="_" format=" ", replaces underscores with spaces.
- The third, pattern="^VMware-" format="", removes any occurrence of VMWare- from the beginning of a string.
<Trim> specifies that leading and trailing white space characters in the naming attribute value must be removed with the trim() method of the Java String class.
<UpperCase> or <LowerCase> specifies that all characters in the naming attribute value must be converted to a particular case. The conversion is done with the toUpperCase() or toLowerCase() method of the Java String class.
Variant searches a database table to normalize different naming attribute values to a single common value. For example, with the <Variant> rule, the naming attribute values“IBM Corp.”, “International Business Machines”, and “IBM” can be normalized into a single value.
This element requires three attributes, which are specified in lowercase:
- table identifies the database table name that is queried.
- variant identifies the column name, from the specified table, which contains the naming attribute input value.
- name identifies the column name, from the specified table, which contains the common output attribute output value that can be mapped from multiple input values.

Feedback