IBM Analytics Solutions Foundation artifacts

IBM® Analytic Solutions Foundation is an application that enables you to define and manage aggregates and build the orchestration process to feed that aggregated data into predictive models. Solutions created with IBM Analytics Solutions Foundation consume events and make recommendations. They use profiles and key performance indicators to aggregate the events. These aggregations are used to provide scores using predictive models. Decision management makes recommendations based on the values of the scores.

A content author can provide a physical model, logical model and orchestration in the pmq_models folder.

Logical model

The logical model file SolutionLogicalModel.XML is found in the pmq_models\logical_definition folder.

A logical model is used to define a logical view of the physical data model. Depending on the business case, an author can define one or many logical models. A content author can define relationships between entities. For example, an author can define hierarchical relationships for various types of equipment such as the following hierarchy:

Equipment
    Electrical equipment
        Transformer
            Distribution Transformer

The following is a metadata representation of the logical model shows how the hierarchy is defined within the logical model. For example, resource is a parent and asset is a child. It also captures asset attributes and their properties such as display name and data type.

<entitySubType code="Asset" description="A resource of type asset" 
displayName="Asset" extends="Resource" id="" version="1.0" author="IBM" >  
    <attribute code="installationDate" displayName="Installation Date"
 dataType="timestamp" description="Installation Date" isRequired="true"/>
    <attribute code="CriticalValue" displayName="Critical Value"
 dataType="double" description="Critical Value of an asset" isRequired="false"/>
    <attributeAlias code="SerialNumber" description="Serial Number of the Asset"
 displayName="Asset Serial Number" attributeRef="ResourceCd1" />
    <attributeAlias code="Model" description="Model of the Asset"
 displayName="Asset Model" attributeRef="ResourceCd2" />
    <attributeAlias code="AssetName" description="Name of the Asset"
 displayName="Asset Name" attributeRef="ResourceName"/>
    <attributeAlias code="AssetLocation" description="Location of the Asset"
 displayName="Asset Location" attributeRef="ResourceLocation"/>
    <attributeAlias code="ParentSerialNumber"
 displayName="Parent Asset Serial Number"
 description="Serial Number of the parent asset" attributeRef="ParentResourceCd1" />
    <attributeAlias code="ParentModel"
 displayName="Parent Asset Model"
 description="Model of the parent asset" attributeRef="ParentResourceCd2" />    
</entitySubType>
  <entitySubType code="Equipment" displayName="Equipment"
 description="An asset of type equipment" extends="Asset" id=""
 version="1.0" author="IBM" >    
  </entitySubType>
<entitySubType code="Machine" displayName="Machine"
 description="Sample Machine" extends="Asset" id="" version="1.0" author="IBM" >
  	 <attribute code="LoadRating" displayName="Load Rating"
 dataType="double" description="Load Rating" isRequired="false"/>   
  	 <attribute code="MTBF" displayName="Mean Time Between Failures"
 dataType="double" description="Mean Time Between Failures" isRequired="false"/>
  </entitySubType>

Similarly a content author can define various events and event subtypes. This logical model contains the definition for a maintenance event that contains details about its attributes and event sub types depending on type of equipment. For example a Machine provides various reading such as “Temperature”, “Ambient Temperature” and so on. The event sub type definition for Machine can be defined as follows:

<eventSubType code="MachineEvent" description="Represents a Machine Event"
 displayName="Machine Event" extends="MaintenanceEvent" id="" version="1.0" author="IBM" >
    <typePropertyValue propertyRef="ResourceType" value="Machine" />    
	<attribute code="Temperature" displayName="Temperature"
 dataType="double" description="Temperature"/>
	<attribute code="AmbientTemperature" displayName="Ambient Temperature"
 dataType="double" description="Ambient Temperature"/>
	<attribute code="Load" displayName="Load"
 dataType="double" description="Load"/>
	<attribute code="Vibration" displayName="Vibration"
 dataType="double" description="Vibration"/>
	<attribute code="Exception" displayName="Exception"
 dataType="double" description="Exception"/>
	<attribute code="Overload" displayName="Overload"
 dataType="double" description="Overload"/>
	<attribute code="CumulativeLoad" displayName="Cumulative Load"
 dataType="double" description="Cumulative Load"/>
	<attribute code="CumulativeOverload" displayName="Cumulative Overload"
 dataType="double" description="Cumulative Overload"/>
	<attribute code="TemperatureRatio" displayName="Temperature Ratio"
 dataType="double" description="Temperature Ratio"/>
	<attribute code="CumulativeStoppage" displayName="Cumulative Stoppage"
 dataType="double" description="Cumulative Stoppage"/>	
 </eventSubType> 

Solution definition

The solution definition file PMQ_solution_definition.xml is found in the pmq_models\solution_definition folder. This solution configuration contains definition for master tables, event tables and profile tables. It also contains definition for services, calculation referred during orchestration.

A content author can provide solution configuration XML files in this folder. A solution configuration XML file is used to define the structure of a physical data model, various service definition, and calculation definition that are referred during orchestrations. For any solution, there is only one solution definition.

Orchestration definitions

The orchestration definition files are found in the pmq_models\orchestration_definition folder. A content author can add orchestrations to this folder. For example, a content author can add an orchestration to process raw data and generate aggregate data, to invoke the predictive model, or to process model scoring results.

The orchestrations provided with the default content pack are described in the following table:

Table 1. Orchestration definitions
Orchestration definition file name Description
PMQ_orchestration_definition_FAS.xml

Contains rules to use when events such as FailureEvent, StoppageEvent and AlertEvent are reported.

PMQ_orchestration_definition_job.xml

Contains rules to use when a user clicks Go on the Analyze Data tab.

PMQ_orchestration_definition_jobscore.xml

Not used at this time.

PMQ_orchestration_definition_jobtest.xml

Contains rules to use when user click Retest on the Test Results tab.

PMQ_orchestration_definition_maintenance.xml

Contains rules to use when a user sends maintenance data.

PMQ_orchestration_definition_poc.xml

Contains sample orchestrations that a content author can use as examples when creating their own orchestrations.

PMQ_orchestration_definition_recommended.xml

Contains orchestrations executed when predictive model output results are processed.

PMQ_orchestration_definition_recommended.xml

Contains rules to be used for the sample Machine.

PMQ_orchestration_definition_sample_recommended.xml

Contains rules that specify how sample predictive models results are processed.

PMQ_orchestration_definition_topn.xml

Contains rules that specify how TopN model output results are processed.